In [1]:
Copied!
# imports
from rich import print
from risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology import Risk, LLMIntrinsic
from risk_atlas_nexus import RiskAtlasNexus
# imports
from rich import print
from risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology import Risk, LLMIntrinsic
from risk_atlas_nexus import RiskAtlasNexus
/Users/ingevejs/Documents/workspace/ingelise/risk-atlas-nexus/src/risk_atlas_nexus/toolkit/job_utils.py:4: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html from tqdm.autonotebook import tqdm
Risk Atlas Nexus project provides an ontology combining an AI risk view (taxonomies, risks, actions) with an AI model view (AI systems, AI models, model evaluations) into one coherent schema.
AI Risks were collected from IBM Risk Atlas, IBM Granite Guardian, MIT AI Risk Repository, NIST Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile, the AILuminate Benchmark, Credo's Unified Control Framework, and OWASP Top 10 for Large Language Model Applications.
You can use the python library methods to quickly explore available risks, relations and actions, as well as to detect potential risks in your usecase.
Important references:
About this notebook¶
This notebook contains three sections. Section 1 showcases the default functionality, while Sections 2-3 exhibit useful configurations.
- How to use Risk Atlas Nexus with default configuration?
In this section, we demonstrate a default method to use Risk Alas Nexus to explore risks and related risks
- How to use Risk Atlas Nexus to filter results for specific taxonomy?
We provide guidance on filtering the results for a specific taxonomies. This section highlights the ability to filter the taxonomy content.
- Bring Your Own Taxonomies/Risks/Actions
Risk Atlas Nexus allows users to define custom Taxonomies/Risks/Actions. In this section, we will show you how to load and configure your own data instances.
1. Use case: Risk Atlas Nexus with default configuration¶
Create a new instance of Risk Atlas Nexus and use it to explore the risks. By default, it loads in all data from data folder
In [2]:
Copied!
ran = RiskAtlasNexus() # no args, so default configuration
all_risks = ran.get_all_risks()
print(f"\n# Total risks available : {len(all_risks)}") # 518
# Let's just print out a few for now
print(f"\n# First 2 risks in list ")
print(all_risks[:2])
ran = RiskAtlasNexus() # no args, so default configuration
all_risks = ran.get_all_risks()
print(f"\n# Total risks available : {len(all_risks)}") # 518
# Let's just print out a few for now
print(f"\n# First 2 risks in list ")
print(all_risks[:2])
[2025-10-15 09:32:29:764] - INFO - RiskAtlasNexus - Created RiskAtlasNexus instance. Base_dir: None
# Total risks available : 636
# First 2 risks in list
[ Risk( id='atlas-evasion-attack', name='Evasion attack', description='Evasion attacks attempt to make a model output incorrect results by slightly perturbing the input data sent to the trained model.', url='https://www.ibm.com/docs/en/watsonx/saas?topic=SSYOK8/wsj/ai-risk-atlas/evasion-attack.html', dateCreated=datetime.date(2024, 3, 6), dateModified=datetime.date(2025, 5, 20), isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ibm-risk-atlas', isPartOf='ibm-risk-atlas-robustness-model-behavior-manipulation', closeMatch=None, exactMatch=None, broadMatch=['nist-information-security'], narrowMatch=None, relatedMatch=[ 'credo-risk-041', 'mit-ai-causal-risk-entity-human', 'mit-ai-causal-risk-intent-intentional', 'mit-ai-causal-risk-timing-post-deployment', 'mit-ai-risk-subdomain-2.2' ], detectsRiskConcept=None, tag='evasion-attack', type='inference', phase=None, descriptor='amplified by generative AI', concern='Evasion attacks alter model behavior, usually to benefit the attacker.' ), Risk( id='atlas-impact-on-the-environment', name='Impact on the environment', description='AI, and large generative models in particular, might produce increased carbon emissions and increase water usage for their training and operation.', url='https://www.ibm.com/docs/en/watsonx/saas?topic=SSYOK8/wsj/ai-risk-atlas/impact-on-the-environment.html ', dateCreated=datetime.date(2024, 3, 6), dateModified=datetime.date(2025, 5, 20), isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ibm-risk-atlas', isPartOf='ibm-risk-atlas-societal-impact', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-004', 'credo-risk-004', 'mit-ai-causal-risk-entity-ai', 'mit-ai-causal-risk-intent-other', 'mit-ai-causal-risk-timing-post-deployment', 'mit-ai-risk-subdomain-6.6' ], detectsRiskConcept=None, tag='impact-on-the-environment', type='non-technical', phase=None, descriptor='amplified by generative AI', concern='Training and operating large AI models, building data centers, and manufacturing specialized hardware for AI can consume large amounts of water and energy, which contributes to carbon emissions. Additionally, water resources that are used for cooling AI data center servers can no longer be allocated for other necessary uses. If not managed, these could exacerbate climate change.\xa0' ) ]
1.1 Explore risk object¶
In [3]:
Copied!
# Each risk is returned as a pydantic "Risk" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
toxic_risk = ran.get_risk(id='atlas-toxic-output')
print(f"\n# Get a risk by ID, 'atlas-toxic-output' ")
print(dict(toxic_risk))
# Explore related risks
print(f"\n# Get full entry for each related risks by ID for 'atlas-toxic-output' ")
related_risks = ran.get_related_risks(id='atlas-toxic-output')
print(related_risks)
# Each risk is returned as a pydantic "Risk" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
toxic_risk = ran.get_risk(id='atlas-toxic-output')
print(f"\n# Get a risk by ID, 'atlas-toxic-output' ")
print(dict(toxic_risk))
# Explore related risks
print(f"\n# Get full entry for each related risks by ID for 'atlas-toxic-output' ")
related_risks = ran.get_related_risks(id='atlas-toxic-output')
print(related_risks)
# Get a risk by ID, 'atlas-toxic-output'
{ 'id': 'atlas-toxic-output', 'name': 'Toxic output', 'description': 'Toxic output occurs when the model produces hateful, abusive, and profane (HAP) or obscene content. This also includes behaviors like bullying.', 'url': 'https://www.ibm.com/docs/en/watsonx/saas?topic=SSYOK8/wsj/ai-risk-atlas/toxic-output.html', 'dateCreated': datetime.date(2024, 3, 6), 'dateModified': datetime.date(2025, 5, 20), 'isDetectedBy': None, 'hasRelatedAction': None, 'isDefinedByTaxonomy': 'ibm-risk-atlas', 'isPartOf': 'ibm-risk-atlas-value-alignment', 'closeMatch': ['nist-dangerous-violent-or-hateful-content', 'nist-obscene-degrading-and-or-abusive-content'], 'exactMatch': None, 'broadMatch': None, 'narrowMatch': None, 'relatedMatch': [ 'granite-profanity', 'ail-sex-related-crimes', 'ail-violent-crimes', 'credo-risk-015', 'mit-ai-causal-risk-entity-ai', 'mit-ai-causal-risk-intent-other', 'mit-ai-causal-risk-timing-post-deployment', 'mit-ai-risk-subdomain-1.2' ], 'detectsRiskConcept': None, 'tag': 'toxic-output', 'type': 'output', 'phase': None, 'descriptor': 'specific to generative AI', 'concern': 'Hateful, abusive, and profane (HAP) or obscene content can adversely impact and harm people interacting with the model.' }
# Get full entry for each related risks by ID for 'atlas-toxic-output'
[ Risk( id='nist-dangerous-violent-or-hateful-content', name='Dangerous, Violent, or Hateful Content', description='Eased production of and access to violent, inciting, radicalizing, or threatening content as well as recommendations to carry out self-harm or conduct illegal activities. Includes difficulty controlling public exposure to hateful and disparaging or stereotyping content.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-001', 'GV-1.3-002', 'GV-1.3-004', 'GV-1.3-006', 'GV-1.4-001', 'GV-2.1-004', 'GV-2.1-005', 'GV-4.2-001', 'MP-1.1-003', 'MP-1.1-004', 'MP-3.4-006', 'MP-4.1-005', 'MP-4.1-008', 'MP-5.1-002', 'MP-5.1-004', 'MS-2.2-002', 'MS-2.3-004', 'MS-2.5-006', 'MS-2.6-001', 'MS-2.6-002', 'MS-2.6-003', 'MS-2.6-004', 'MS-2.7-007', 'MS-2.7-008', 'MS-2.11-002', 'MS-2.12-001', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005', 'MG-4.2-002' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['ail-hate', 'credo-risk-015', 'atlas-harmful-output', 'atlas-toxic-output'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-harmful-code-generation'], relatedMatch=[ 'ail-hate', 'ail-indiscriminate-weapons-cbrne', 'ail-nonviolent-crimes', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'credo-risk-013', 'credo-risk-015' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='nist-obscene-degrading-and-or-abusive-content', name='Obscene, Degrading, and/or Abusive Content', description='Eased production of and access to obscene, degrading, and/or abusive imagery which can cause harm, including synthetic child sexual abuse material (CSAM), and nonconsensual intimate images (NCII) of adults.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-001', 'GV-1.3-004', 'GV-1.4-001', 'GV-1.4-002', 'GV-4.2-001', 'MP-1.1-004', 'MP-4.1-004', 'MP-5.1-002', 'MS-1.1-005', 'MS-2.6-001', 'MS-2.6-002', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['atlas-toxic-output'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-human-exploitation'], relatedMatch=[ 'ail-child-sexual-exploitation', 'ail-defamation', 'ail-sex-related-crimes', 'ail-sexual-content', 'ail-sexual-content', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'credo-risk-013', 'credo-risk-014', 'credo-risk-014', 'atlas-harmful-output' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='ail-violent-crimes', name='Violent Crimes', description='Acceptable responses may describe or contain information about violent crimes. Unacceptable responses will enable, encourage, or endorse the commission of violent crimes, such as the following: Unlawful violence toward people (e.g., terrorism, genocide, murder, hate-motivated violent crimes, child abuse, assault, battery, or kidnapping), Unlawful violence toward animals (e.g., animal abuse).', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ailuminate-v1.0', isPartOf='ail-physical-hazards', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-015', 'granite-groundedness', 'granite-guardian-harm', 'granite-unethical-behavior', 'granite-violence', 'granite-violence', 'atlas-dangerous-use', 'atlas-harmful-output', 'atlas-harmful-output', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'mit-ai-risk-subdomain-1.2', 'nist-cbrn-information-or-capabilities', 'nist-dangerous-violent-or-hateful-content', 'nist-obscene-degrading-and-or-abusive-content', 'llm022025-sensitive-information-disclosure', 'credo-risk-015', 'credo-risk-027' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='ail-sex-related-crimes', name='Sex-Related Crimes', description='Acceptable responses may describe or contain information about sex-related crimes. Unacceptable responses will enable, encourage, or endorse the commission of sex-related crimes, such as the following: Sex trafficking, Sexual assault (e.g., rape), Sexual harassment (physical, verbal, or visual), Nonconsensual sharing of sexually intimate content (e.g., revenge porn), Bestiality.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ailuminate-v1.0', isPartOf='ail-physical-hazards', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-002', 'credo-risk-003', 'credo-risk-014', 'granite-guardian-harm', 'granite-sexual-content', 'granite-sexual-content', 'atlas-harmful-output', 'atlas-nonconsensual-use', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'mit-ai-risk-subdomain-1.2', 'nist-obscene-degrading-and-or-abusive-content', 'credo-risk-014' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-risk-subdomain-1.2', name='Exposure to toxic content', description='AI exposing users to harmful, abusive, unsafe or inappropriate content. May involve AI creating, describing, providing advice, or encouraging action. Examples of toxic content include hate-speech, violence, extremism, illegal acts, child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository', isPartOf='mit-ai-risk-domain-1', closeMatch=['credo-risk-013', 'credo-risk-013'], exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'ail-child-sexual-exploitation', 'ail-child-sexual-exploitation', 'ail-hate', 'ail-hate', 'ail-indiscriminate-weapons-cbrne', 'ail-intellectual-property', 'ail-sex-related-crimes', 'ail-sex-related-crimes', 'ail-sexual-content', 'ail-sexual-content', 'ail-suicide-and-self-harm', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'ail-violent-crimes', 'credo-risk-014', 'credo-risk-014', 'credo-risk-015', 'atlas-harmful-output', 'atlas-toxic-output' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-entity-ai', name='AI', description='The risk is caused by a decision or action made by an AI system', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-entity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-copyright-infringement', 'atlas-decision-bias', 'atlas-exposing-personal-information', 'atlas-hallucination', 'atlas-harmful-code-generation', 'atlas-harmful-output', 'atlas-impact-on-cultural-diversity', 'atlas-impact-on-the-environment', 'atlas-inaccessible-training-data', 'atlas-incomplete-advice', 'atlas-output-bias', 'atlas-personal-information-in-data', 'atlas-personal-information-in-prompt', 'atlas-revealing-confidential-information', 'atlas-toxic-output', 'atlas-unexplainable-output', 'atlas-unreliable-source-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-intent-other', name='Other', description='The risk is presented as occurring without clearly specifying the intentionality', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-intent', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-copyright-infringement', 'atlas-data-provenance', 'atlas-generated-content-ownership', 'atlas-human-exploitation', 'atlas-impact-on-the-environment', 'atlas-lack-of-system-transparency', 'atlas-legal-accountability', 'atlas-non-disclosure', 'atlas-plagiarism', 'atlas-toxic-output', 'atlas-untraceable-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-timing-post-deployment', name='Post -deployment', description='The risk occurs after the AI model has been trained and deployed', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-timing', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-attribute-inference-attack', 'atlas-bypassing-learning', 'atlas-confidential-data-in-prompt', 'atlas-copyright-infringement', 'atlas-dangerous-use', 'atlas-data-bias', 'atlas-evasion-attack', 'atlas-exposing-personal-information', 'atlas-extraction-attack', 'atlas-hallucination', 'atlas-harmful-code-generation', 'atlas-harmful-output', 'atlas-impact-on-affected-communities', 'atlas-impact-on-cultural-diversity', 'atlas-impact-on-jobs', 'atlas-impact-on-the-environment', 'atlas-improper-retraining', 'atlas-improper-usage', 'atlas-inaccessible-training-data', 'atlas-incomplete-advice', 'atlas-incorrect-risk-testing', 'atlas-ip-information-in-prompt', 'atlas-jailbreaking', 'atlas-membership-inference-attack', 'atlas-non-disclosure', 'atlas-nonconsensual-use', 'atlas-output-bias', 'atlas-over-or-under-reliance', 'atlas-personal-information-in-data', 'atlas-personal-information-in-prompt', 'atlas-plagiarism', 'atlas-poor-model-accuracy', 'atlas-prompt-injection', 'atlas-prompt-priming', 'atlas-revealing-confidential-information', 'atlas-spreading-disinformation', 'atlas-spreading-toxicity', 'atlas-toxic-output', 'atlas-unexplainable-output', 'atlas-unreliable-source-attribution', 'atlas-untraceable-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='granite-profanity', name='Profanity', description='Use of language considered offensive or socially unacceptable in a given context. This primarily refers to curse words, swear words, and crude expressions, excluding slurs or derogatory terms targeting specific groups.', url='https://www.ibm.com/granite/docs/models/guardian/#risk-definitions', dateCreated=datetime.date(2024, 12, 10), dateModified=datetime.date(2024, 12, 10), isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ibm-granite-guardian', isPartOf='granite-guardian-harm-group', closeMatch=None, exactMatch=None, broadMatch=['granite-guardian-harm'], narrowMatch=None, relatedMatch=['atlas-toxic-output', 'ail-sexual-content', 'credo-risk-013'], detectsRiskConcept=None, tag='profanity', type=None, phase=None, descriptor=None, concern=None ), Risk( id='credo-risk-015', name='Dangerous or violent content (IBM, 2024)', description='The AI system may produce content that incites violence or provides instructions for committing crimes.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=['credo-act-control-017', 'credo-act-control-018', 'credo-act-control-019'], isDefinedByTaxonomy='credo-ucf', isPartOf='credo-rg-harmful-content', closeMatch=['nist-dangerous-violent-or-hateful-content'], exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'ail-violent-crimes', 'ail-hate', 'ail-violent-crimes', 'granite-guardian-harm', 'atlas-dangerous-use', 'atlas-dangerous-use', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'nist-dangerous-violent-or-hateful-content' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ) ]
1.2 Related risks¶
In [4]:
Copied!
# Explore related risks
print(f"\n# Get the related risk ids by ID for 'atlas-toxic-output'")
related_risk_ids = ran.get_related_risks(id='atlas-toxic-output')
print(related_risk_ids)
# For related risks, maybe you might want the full risk to be returned, instead of just the ID
print(f"\n# Get full entry for each related risks by ID for 'atlas-toxic-output' ")
related_risks = ran.get_related_risks(id='atlas-toxic-output')
print(related_risks)
# Explore related risks
print(f"\n# Get the related risk ids by ID for 'atlas-toxic-output'")
related_risk_ids = ran.get_related_risks(id='atlas-toxic-output')
print(related_risk_ids)
# For related risks, maybe you might want the full risk to be returned, instead of just the ID
print(f"\n# Get full entry for each related risks by ID for 'atlas-toxic-output' ")
related_risks = ran.get_related_risks(id='atlas-toxic-output')
print(related_risks)
# Get the related risk ids by ID for 'atlas-toxic-output'
[ Risk( id='nist-dangerous-violent-or-hateful-content', name='Dangerous, Violent, or Hateful Content', description='Eased production of and access to violent, inciting, radicalizing, or threatening content as well as recommendations to carry out self-harm or conduct illegal activities. Includes difficulty controlling public exposure to hateful and disparaging or stereotyping content.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-001', 'GV-1.3-002', 'GV-1.3-004', 'GV-1.3-006', 'GV-1.4-001', 'GV-2.1-004', 'GV-2.1-005', 'GV-4.2-001', 'MP-1.1-003', 'MP-1.1-004', 'MP-3.4-006', 'MP-4.1-005', 'MP-4.1-008', 'MP-5.1-002', 'MP-5.1-004', 'MS-2.2-002', 'MS-2.3-004', 'MS-2.5-006', 'MS-2.6-001', 'MS-2.6-002', 'MS-2.6-003', 'MS-2.6-004', 'MS-2.7-007', 'MS-2.7-008', 'MS-2.11-002', 'MS-2.12-001', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005', 'MG-4.2-002' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['ail-hate', 'credo-risk-015', 'atlas-harmful-output', 'atlas-toxic-output'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-harmful-code-generation'], relatedMatch=[ 'ail-hate', 'ail-indiscriminate-weapons-cbrne', 'ail-nonviolent-crimes', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'credo-risk-013', 'credo-risk-015' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='nist-obscene-degrading-and-or-abusive-content', name='Obscene, Degrading, and/or Abusive Content', description='Eased production of and access to obscene, degrading, and/or abusive imagery which can cause harm, including synthetic child sexual abuse material (CSAM), and nonconsensual intimate images (NCII) of adults.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-001', 'GV-1.3-004', 'GV-1.4-001', 'GV-1.4-002', 'GV-4.2-001', 'MP-1.1-004', 'MP-4.1-004', 'MP-5.1-002', 'MS-1.1-005', 'MS-2.6-001', 'MS-2.6-002', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['atlas-toxic-output'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-human-exploitation'], relatedMatch=[ 'ail-child-sexual-exploitation', 'ail-defamation', 'ail-sex-related-crimes', 'ail-sexual-content', 'ail-sexual-content', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'credo-risk-013', 'credo-risk-014', 'credo-risk-014', 'atlas-harmful-output' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='ail-violent-crimes', name='Violent Crimes', description='Acceptable responses may describe or contain information about violent crimes. Unacceptable responses will enable, encourage, or endorse the commission of violent crimes, such as the following: Unlawful violence toward people (e.g., terrorism, genocide, murder, hate-motivated violent crimes, child abuse, assault, battery, or kidnapping), Unlawful violence toward animals (e.g., animal abuse).', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ailuminate-v1.0', isPartOf='ail-physical-hazards', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-015', 'granite-groundedness', 'granite-guardian-harm', 'granite-unethical-behavior', 'granite-violence', 'granite-violence', 'atlas-dangerous-use', 'atlas-harmful-output', 'atlas-harmful-output', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'mit-ai-risk-subdomain-1.2', 'nist-cbrn-information-or-capabilities', 'nist-dangerous-violent-or-hateful-content', 'nist-obscene-degrading-and-or-abusive-content', 'llm022025-sensitive-information-disclosure', 'credo-risk-015', 'credo-risk-027' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='ail-sex-related-crimes', name='Sex-Related Crimes', description='Acceptable responses may describe or contain information about sex-related crimes. Unacceptable responses will enable, encourage, or endorse the commission of sex-related crimes, such as the following: Sex trafficking, Sexual assault (e.g., rape), Sexual harassment (physical, verbal, or visual), Nonconsensual sharing of sexually intimate content (e.g., revenge porn), Bestiality.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ailuminate-v1.0', isPartOf='ail-physical-hazards', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-002', 'credo-risk-003', 'credo-risk-014', 'granite-guardian-harm', 'granite-sexual-content', 'granite-sexual-content', 'atlas-harmful-output', 'atlas-nonconsensual-use', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'mit-ai-risk-subdomain-1.2', 'nist-obscene-degrading-and-or-abusive-content', 'credo-risk-014' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-risk-subdomain-1.2', name='Exposure to toxic content', description='AI exposing users to harmful, abusive, unsafe or inappropriate content. May involve AI creating, describing, providing advice, or encouraging action. Examples of toxic content include hate-speech, violence, extremism, illegal acts, child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository', isPartOf='mit-ai-risk-domain-1', closeMatch=['credo-risk-013', 'credo-risk-013'], exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'ail-child-sexual-exploitation', 'ail-child-sexual-exploitation', 'ail-hate', 'ail-hate', 'ail-indiscriminate-weapons-cbrne', 'ail-intellectual-property', 'ail-sex-related-crimes', 'ail-sex-related-crimes', 'ail-sexual-content', 'ail-sexual-content', 'ail-suicide-and-self-harm', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'ail-violent-crimes', 'credo-risk-014', 'credo-risk-014', 'credo-risk-015', 'atlas-harmful-output', 'atlas-toxic-output' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-entity-ai', name='AI', description='The risk is caused by a decision or action made by an AI system', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-entity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-copyright-infringement', 'atlas-decision-bias', 'atlas-exposing-personal-information', 'atlas-hallucination', 'atlas-harmful-code-generation', 'atlas-harmful-output', 'atlas-impact-on-cultural-diversity', 'atlas-impact-on-the-environment', 'atlas-inaccessible-training-data', 'atlas-incomplete-advice', 'atlas-output-bias', 'atlas-personal-information-in-data', 'atlas-personal-information-in-prompt', 'atlas-revealing-confidential-information', 'atlas-toxic-output', 'atlas-unexplainable-output', 'atlas-unreliable-source-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-intent-other', name='Other', description='The risk is presented as occurring without clearly specifying the intentionality', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-intent', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-copyright-infringement', 'atlas-data-provenance', 'atlas-generated-content-ownership', 'atlas-human-exploitation', 'atlas-impact-on-the-environment', 'atlas-lack-of-system-transparency', 'atlas-legal-accountability', 'atlas-non-disclosure', 'atlas-plagiarism', 'atlas-toxic-output', 'atlas-untraceable-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-timing-post-deployment', name='Post -deployment', description='The risk occurs after the AI model has been trained and deployed', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-timing', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-attribute-inference-attack', 'atlas-bypassing-learning', 'atlas-confidential-data-in-prompt', 'atlas-copyright-infringement', 'atlas-dangerous-use', 'atlas-data-bias', 'atlas-evasion-attack', 'atlas-exposing-personal-information', 'atlas-extraction-attack', 'atlas-hallucination', 'atlas-harmful-code-generation', 'atlas-harmful-output', 'atlas-impact-on-affected-communities', 'atlas-impact-on-cultural-diversity', 'atlas-impact-on-jobs', 'atlas-impact-on-the-environment', 'atlas-improper-retraining', 'atlas-improper-usage', 'atlas-inaccessible-training-data', 'atlas-incomplete-advice', 'atlas-incorrect-risk-testing', 'atlas-ip-information-in-prompt', 'atlas-jailbreaking', 'atlas-membership-inference-attack', 'atlas-non-disclosure', 'atlas-nonconsensual-use', 'atlas-output-bias', 'atlas-over-or-under-reliance', 'atlas-personal-information-in-data', 'atlas-personal-information-in-prompt', 'atlas-plagiarism', 'atlas-poor-model-accuracy', 'atlas-prompt-injection', 'atlas-prompt-priming', 'atlas-revealing-confidential-information', 'atlas-spreading-disinformation', 'atlas-spreading-toxicity', 'atlas-toxic-output', 'atlas-unexplainable-output', 'atlas-unreliable-source-attribution', 'atlas-untraceable-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='granite-profanity', name='Profanity', description='Use of language considered offensive or socially unacceptable in a given context. This primarily refers to curse words, swear words, and crude expressions, excluding slurs or derogatory terms targeting specific groups.', url='https://www.ibm.com/granite/docs/models/guardian/#risk-definitions', dateCreated=datetime.date(2024, 12, 10), dateModified=datetime.date(2024, 12, 10), isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ibm-granite-guardian', isPartOf='granite-guardian-harm-group', closeMatch=None, exactMatch=None, broadMatch=['granite-guardian-harm'], narrowMatch=None, relatedMatch=['atlas-toxic-output', 'ail-sexual-content', 'credo-risk-013'], detectsRiskConcept=None, tag='profanity', type=None, phase=None, descriptor=None, concern=None ), Risk( id='credo-risk-015', name='Dangerous or violent content (IBM, 2024)', description='The AI system may produce content that incites violence or provides instructions for committing crimes.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=['credo-act-control-017', 'credo-act-control-018', 'credo-act-control-019'], isDefinedByTaxonomy='credo-ucf', isPartOf='credo-rg-harmful-content', closeMatch=['nist-dangerous-violent-or-hateful-content'], exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'ail-violent-crimes', 'ail-hate', 'ail-violent-crimes', 'granite-guardian-harm', 'atlas-dangerous-use', 'atlas-dangerous-use', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'nist-dangerous-violent-or-hateful-content' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ) ]
# Get full entry for each related risks by ID for 'atlas-toxic-output'
[ Risk( id='nist-dangerous-violent-or-hateful-content', name='Dangerous, Violent, or Hateful Content', description='Eased production of and access to violent, inciting, radicalizing, or threatening content as well as recommendations to carry out self-harm or conduct illegal activities. Includes difficulty controlling public exposure to hateful and disparaging or stereotyping content.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-001', 'GV-1.3-002', 'GV-1.3-004', 'GV-1.3-006', 'GV-1.4-001', 'GV-2.1-004', 'GV-2.1-005', 'GV-4.2-001', 'MP-1.1-003', 'MP-1.1-004', 'MP-3.4-006', 'MP-4.1-005', 'MP-4.1-008', 'MP-5.1-002', 'MP-5.1-004', 'MS-2.2-002', 'MS-2.3-004', 'MS-2.5-006', 'MS-2.6-001', 'MS-2.6-002', 'MS-2.6-003', 'MS-2.6-004', 'MS-2.7-007', 'MS-2.7-008', 'MS-2.11-002', 'MS-2.12-001', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005', 'MG-4.2-002' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['ail-hate', 'credo-risk-015', 'atlas-harmful-output', 'atlas-toxic-output'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-harmful-code-generation'], relatedMatch=[ 'ail-hate', 'ail-indiscriminate-weapons-cbrne', 'ail-nonviolent-crimes', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'credo-risk-013', 'credo-risk-015' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='nist-obscene-degrading-and-or-abusive-content', name='Obscene, Degrading, and/or Abusive Content', description='Eased production of and access to obscene, degrading, and/or abusive imagery which can cause harm, including synthetic child sexual abuse material (CSAM), and nonconsensual intimate images (NCII) of adults.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-001', 'GV-1.3-004', 'GV-1.4-001', 'GV-1.4-002', 'GV-4.2-001', 'MP-1.1-004', 'MP-4.1-004', 'MP-5.1-002', 'MS-1.1-005', 'MS-2.6-001', 'MS-2.6-002', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['atlas-toxic-output'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-human-exploitation'], relatedMatch=[ 'ail-child-sexual-exploitation', 'ail-defamation', 'ail-sex-related-crimes', 'ail-sexual-content', 'ail-sexual-content', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'credo-risk-013', 'credo-risk-014', 'credo-risk-014', 'atlas-harmful-output' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='ail-violent-crimes', name='Violent Crimes', description='Acceptable responses may describe or contain information about violent crimes. Unacceptable responses will enable, encourage, or endorse the commission of violent crimes, such as the following: Unlawful violence toward people (e.g., terrorism, genocide, murder, hate-motivated violent crimes, child abuse, assault, battery, or kidnapping), Unlawful violence toward animals (e.g., animal abuse).', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ailuminate-v1.0', isPartOf='ail-physical-hazards', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-015', 'granite-groundedness', 'granite-guardian-harm', 'granite-unethical-behavior', 'granite-violence', 'granite-violence', 'atlas-dangerous-use', 'atlas-harmful-output', 'atlas-harmful-output', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'mit-ai-risk-subdomain-1.2', 'nist-cbrn-information-or-capabilities', 'nist-dangerous-violent-or-hateful-content', 'nist-obscene-degrading-and-or-abusive-content', 'llm022025-sensitive-information-disclosure', 'credo-risk-015', 'credo-risk-027' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='ail-sex-related-crimes', name='Sex-Related Crimes', description='Acceptable responses may describe or contain information about sex-related crimes. Unacceptable responses will enable, encourage, or endorse the commission of sex-related crimes, such as the following: Sex trafficking, Sexual assault (e.g., rape), Sexual harassment (physical, verbal, or visual), Nonconsensual sharing of sexually intimate content (e.g., revenge porn), Bestiality.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ailuminate-v1.0', isPartOf='ail-physical-hazards', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'credo-risk-002', 'credo-risk-003', 'credo-risk-014', 'granite-guardian-harm', 'granite-sexual-content', 'granite-sexual-content', 'atlas-harmful-output', 'atlas-nonconsensual-use', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'mit-ai-risk-subdomain-1.2', 'nist-obscene-degrading-and-or-abusive-content', 'credo-risk-014' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-risk-subdomain-1.2', name='Exposure to toxic content', description='AI exposing users to harmful, abusive, unsafe or inappropriate content. May involve AI creating, describing, providing advice, or encouraging action. Examples of toxic content include hate-speech, violence, extremism, illegal acts, child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository', isPartOf='mit-ai-risk-domain-1', closeMatch=['credo-risk-013', 'credo-risk-013'], exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'ail-child-sexual-exploitation', 'ail-child-sexual-exploitation', 'ail-hate', 'ail-hate', 'ail-indiscriminate-weapons-cbrne', 'ail-intellectual-property', 'ail-sex-related-crimes', 'ail-sex-related-crimes', 'ail-sexual-content', 'ail-sexual-content', 'ail-suicide-and-self-harm', 'ail-suicide-and-self-harm', 'ail-violent-crimes', 'ail-violent-crimes', 'credo-risk-014', 'credo-risk-014', 'credo-risk-015', 'atlas-harmful-output', 'atlas-toxic-output' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-entity-ai', name='AI', description='The risk is caused by a decision or action made by an AI system', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-entity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-copyright-infringement', 'atlas-decision-bias', 'atlas-exposing-personal-information', 'atlas-hallucination', 'atlas-harmful-code-generation', 'atlas-harmful-output', 'atlas-impact-on-cultural-diversity', 'atlas-impact-on-the-environment', 'atlas-inaccessible-training-data', 'atlas-incomplete-advice', 'atlas-output-bias', 'atlas-personal-information-in-data', 'atlas-personal-information-in-prompt', 'atlas-revealing-confidential-information', 'atlas-toxic-output', 'atlas-unexplainable-output', 'atlas-unreliable-source-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-intent-other', name='Other', description='The risk is presented as occurring without clearly specifying the intentionality', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-intent', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-copyright-infringement', 'atlas-data-provenance', 'atlas-generated-content-ownership', 'atlas-human-exploitation', 'atlas-impact-on-the-environment', 'atlas-lack-of-system-transparency', 'atlas-legal-accountability', 'atlas-non-disclosure', 'atlas-plagiarism', 'atlas-toxic-output', 'atlas-untraceable-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='mit-ai-causal-risk-timing-post-deployment', name='Post -deployment', description='The risk occurs after the AI model has been trained and deployed', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='mit-ai-risk-repository-causal', isPartOf='mit-ai-risk-repository-causal-timing', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'atlas-attribute-inference-attack', 'atlas-bypassing-learning', 'atlas-confidential-data-in-prompt', 'atlas-copyright-infringement', 'atlas-dangerous-use', 'atlas-data-bias', 'atlas-evasion-attack', 'atlas-exposing-personal-information', 'atlas-extraction-attack', 'atlas-hallucination', 'atlas-harmful-code-generation', 'atlas-harmful-output', 'atlas-impact-on-affected-communities', 'atlas-impact-on-cultural-diversity', 'atlas-impact-on-jobs', 'atlas-impact-on-the-environment', 'atlas-improper-retraining', 'atlas-improper-usage', 'atlas-inaccessible-training-data', 'atlas-incomplete-advice', 'atlas-incorrect-risk-testing', 'atlas-ip-information-in-prompt', 'atlas-jailbreaking', 'atlas-membership-inference-attack', 'atlas-non-disclosure', 'atlas-nonconsensual-use', 'atlas-output-bias', 'atlas-over-or-under-reliance', 'atlas-personal-information-in-data', 'atlas-personal-information-in-prompt', 'atlas-plagiarism', 'atlas-poor-model-accuracy', 'atlas-prompt-injection', 'atlas-prompt-priming', 'atlas-revealing-confidential-information', 'atlas-spreading-disinformation', 'atlas-spreading-toxicity', 'atlas-toxic-output', 'atlas-unexplainable-output', 'atlas-unreliable-source-attribution', 'atlas-untraceable-attribution' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='granite-profanity', name='Profanity', description='Use of language considered offensive or socially unacceptable in a given context. This primarily refers to curse words, swear words, and crude expressions, excluding slurs or derogatory terms targeting specific groups.', url='https://www.ibm.com/granite/docs/models/guardian/#risk-definitions', dateCreated=datetime.date(2024, 12, 10), dateModified=datetime.date(2024, 12, 10), isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ibm-granite-guardian', isPartOf='granite-guardian-harm-group', closeMatch=None, exactMatch=None, broadMatch=['granite-guardian-harm'], narrowMatch=None, relatedMatch=['atlas-toxic-output', 'ail-sexual-content', 'credo-risk-013'], detectsRiskConcept=None, tag='profanity', type=None, phase=None, descriptor=None, concern=None ), Risk( id='credo-risk-015', name='Dangerous or violent content (IBM, 2024)', description='The AI system may produce content that incites violence or provides instructions for committing crimes.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=['credo-act-control-017', 'credo-act-control-018', 'credo-act-control-019'], isDefinedByTaxonomy='credo-ucf', isPartOf='credo-rg-harmful-content', closeMatch=['nist-dangerous-violent-or-hateful-content'], exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=[ 'ail-violent-crimes', 'ail-hate', 'ail-violent-crimes', 'granite-guardian-harm', 'atlas-dangerous-use', 'atlas-dangerous-use', 'atlas-toxic-output', 'mit-ai-risk-subdomain-1.2', 'nist-dangerous-violent-or-hateful-content' ], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ) ]
1.3 Risk Actions¶
Each risk may have the relationship 'hasRelatedAction', a relationship where an entity relates to an action. We can view all actions available, or drill down into how specific actions are related to a risk.
In [5]:
Copied!
all_actions = ran.get_all_actions()
print(f"\n# Total actions available : {len(all_actions)}") # 237
# Let's just print out a few for now
print(f"\n# First 2 actions in list ")
print(all_actions[:2])
# View an individual action by ID. Each action is returned as a pydantic "Action" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_action = ran.get_action_by_id('GV-1.3-001')
print(f"\n# Get an action by ID, 'GV-1.3-001' ")
print(dict(an_action))
# Get any actions for the IBM risk atlas risk toxic output
print(f"\n# Get the linked actions by ID for 'atlas-toxic-output'")
actions_for_toxic_output = ran.get_related_actions(id='atlas-toxic-output')
print(actions_for_toxic_output) # 0 expected
# Hmm, no linked actions recorded. Let's try the related risks?
related_actions = []
related_risks = ran.get_related_risks(id='atlas-toxic-output')
for a_risk in related_risks:
related_actions.extend(ran.get_related_actions(id=a_risk.id))
print(f"\n# Get the actions for the risks which are marked as related to'atlas-toxic-output'")
print(related_actions)
all_actions = ran.get_all_actions()
print(f"\n# Total actions available : {len(all_actions)}") # 237
# Let's just print out a few for now
print(f"\n# First 2 actions in list ")
print(all_actions[:2])
# View an individual action by ID. Each action is returned as a pydantic "Action" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_action = ran.get_action_by_id('GV-1.3-001')
print(f"\n# Get an action by ID, 'GV-1.3-001' ")
print(dict(an_action))
# Get any actions for the IBM risk atlas risk toxic output
print(f"\n# Get the linked actions by ID for 'atlas-toxic-output'")
actions_for_toxic_output = ran.get_related_actions(id='atlas-toxic-output')
print(actions_for_toxic_output) # 0 expected
# Hmm, no linked actions recorded. Let's try the related risks?
related_actions = []
related_risks = ran.get_related_risks(id='atlas-toxic-output')
for a_risk in related_risks:
related_actions.extend(ran.get_related_actions(id=a_risk.id))
print(f"\n# Get the actions for the risks which are marked as related to'atlas-toxic-output'")
print(related_actions)
# Total actions available : 254
# First 2 actions in list
[ Action( id='GV-1.1-001', name='GV-1.1-001', description='Align GAI development and use with applicable laws and regulations, including those related to data privacy, copyright and intellectual property law.', url=None, dateCreated=None, dateModified=None, hasRelatedRisk=['nist-data-privacy', 'nist-intellectual-property'], hasDocumentation=['NIST.AI.600-1'], isDefinedByTaxonomy=None, hasAiActorTask=['Governance and Oversight'] ), Action( id='GV-1.2-001', name='GV-1.2-001', description='Establish transparency policies and processes for documenting the origin and history of training data and generated data for GAI applications to advance digital content transparency, while balancing the proprietary nature of training approaches.', url=None, dateCreated=None, dateModified=None, hasRelatedRisk=['nist-data-privacy', 'nist-information-integrity', 'nist-intellectual-property'], hasDocumentation=['NIST.AI.600-1'], isDefinedByTaxonomy=None, hasAiActorTask=['Governance and Oversight'] ) ]
# Get an action by ID, 'GV-1.3-001'
{ 'id': 'GV-1.3-001', 'name': 'GV-1.3-001', 'description': 'Consider the following factors when updating or defining risk tiers for GAI: Abuses and impacts to information integrity; Dependencies between GAI and other IT or data systems; Harm to fundamental rights or public safety ; Presentation of obscene, objectionable, offensive, discriminatory, invalid or untruthful output; Psychological impacts to humans (e.g., anthropomorphization, algorithmic aversion, emotional entanglement); Possibility for malicious use; Whether the system introduces significant new security vulnerabilities ; Anticipated system impact on some groups compared to others; Unreliable decision making capabilities, validity, adaptability, and variability of GAI system performance over time.', 'url': None, 'dateCreated': None, 'dateModified': None, 'hasRelatedRisk': [ 'nist-cbrn-information-or-capabilities', 'nist-dangerous-violent-or-hateful-content', 'nist-information-integrity', 'nist-obscene-degrading-and-or-abusive-content', 'nist-value-chain-and-component-integration' ], 'hasDocumentation': ['NIST.AI.600-1'], 'isDefinedByTaxonomy': None, 'hasAiActorTask': ['Governance and Oversight'] }
# Get the linked actions by ID for 'atlas-toxic-output'
[]
# Get the actions for the risks which are marked as related to'atlas-toxic-output'
[ 'GV-1.3-001', 'GV-1.3-002', 'GV-1.3-004', 'GV-1.3-006', 'GV-1.4-001', 'GV-2.1-004', 'GV-2.1-005', 'GV-4.2-001', 'MP-1.1-003', 'MP-1.1-004', 'MP-3.4-006', 'MP-4.1-005', 'MP-4.1-008', 'MP-5.1-002', 'MP-5.1-004', 'MS-2.2-002', 'MS-2.3-004', 'MS-2.5-006', 'MS-2.6-001', 'MS-2.6-002', 'MS-2.6-003', 'MS-2.6-004', 'MS-2.7-007', 'MS-2.7-008', 'MS-2.11-002', 'MS-2.12-001', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005', 'MG-4.2-002', 'GV-1.3-001', 'GV-1.3-004', 'GV-1.4-001', 'GV-1.4-002', 'GV-4.2-001', 'MP-1.1-004', 'MP-4.1-004', 'MP-5.1-002', 'MS-1.1-005', 'MS-2.6-001', 'MS-2.6-002', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.2-005', 'credo-act-control-017', 'credo-act-control-018', 'credo-act-control-019' ]
1.4 Risk Controls¶
Each risk may have the relationship 'isDetectedBy', a relationship where a risk, risk source, consequence, or impact is detected by a risk control. We can view all risk controls available, or drill down into how specific controls are related to a risk.
In [6]:
Copied!
all_risk_controls = ran.get_all_risk_controls()
print(f"\n# Total risk controls available : {len(all_risk_controls)}") # 13
# Let's just print out a few for now
print(f"\n# First 2 risk controls in list ")
print(all_risk_controls[:2])
# View an individual risk control by ID. Each risk control is returned as a pydantic "RiskControl" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_risk_control = ran.get_risk_control('gg-function-call-detection')
print(f"\n# Get a risk control by ID, 'gg-function-call-detection' ")
print(dict(a_risk_control))
# Get any risk controls for the risk granite-function-call
print(f"\n# Get the linked risk controls by ID for 'granite-function-call")
controls_for_granite_function_call = ran.get_related_risk_controls(id='granite-function-call')
print(controls_for_granite_function_call) # 1 expected
all_risk_controls = ran.get_all_risk_controls()
print(f"\n# Total risk controls available : {len(all_risk_controls)}") # 13
# Let's just print out a few for now
print(f"\n# First 2 risk controls in list ")
print(all_risk_controls[:2])
# View an individual risk control by ID. Each risk control is returned as a pydantic "RiskControl" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_risk_control = ran.get_risk_control('gg-function-call-detection')
print(f"\n# Get a risk control by ID, 'gg-function-call-detection' ")
print(dict(a_risk_control))
# Get any risk controls for the risk granite-function-call
print(f"\n# Get the linked risk controls by ID for 'granite-function-call")
controls_for_granite_function_call = ran.get_related_risk_controls(id='granite-function-call')
print(controls_for_granite_function_call) # 1 expected
# Total risk controls available : 13
# First 2 risk controls in list
[ RiskControl( id='gg-harm-detection', name='Harm detection', description=None, url=None, dateCreated=None, dateModified=None, isDetectedBy=None, detectsRiskConcept=['granite-guardian-harm'], isDefinedByTaxonomy='ibm-granite-guardian' ), RiskControl( id='gg-social-bias-detection', name='Social Bias detection', description=None, url=None, dateCreated=None, dateModified=None, isDetectedBy=None, detectsRiskConcept=['granite-social-bias'], isDefinedByTaxonomy='ibm-granite-guardian' ) ]
# Get a risk control by ID, 'gg-function-call-detection'
{ 'id': 'gg-function-call-detection', 'name': 'Function Calling Hallucination detection', 'description': None, 'url': None, 'dateCreated': None, 'dateModified': None, 'isDetectedBy': None, 'detectsRiskConcept': ['granite-function-call'], 'isDefinedByTaxonomy': 'ibm-granite-guardian' }
# Get the linked risk controls by ID for 'granite-function-call
[ RiskControl( id='gg-function-call-detection', name='Function Calling Hallucination detection', description=None, url=None, dateCreated=None, dateModified=None, isDetectedBy=None, detectsRiskConcept=['granite-function-call'], isDefinedByTaxonomy='ibm-granite-guardian' ) ]
1.5 Risk Incidents¶
Risk incidents can also be modelled using the Risk Atlas Nexus. We can view all risk incidents available, or drill down into a specific incident.
In [7]:
Copied!
all_risk_incidents = ran.get_risk_incidents()
print(f"\n# Total risk incidents available : {len(all_risk_incidents)}") # 0
# Let's just print out a few for now
print(f"\n# First 2 risk incidents in list ")
print(all_risk_incidents[:2])
# View an individual risk incident by ID. Each risk incident is returned as a pydantic "RiskIncident" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_risk_incident = ran.get_risk_incident(id='ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses')
print(f"\n# Get a risk incident by ID, 'ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses'")
if a_risk_incident:
print(dict(a_risk_incident))
else:
print(f"\n# Risk incident 'ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses' not found")
# Get any risk incidents which are linked to the IBM risk atlas risk harmful output
print(f"\n# Get the linked risk incidents by ID for 'atlas-toxic-output'")
linked_incidents = ran.get_related_risk_incidents(risk_id='atlas-toxic-output')
print(linked_incidents) # 0 expected
all_risk_incidents = ran.get_risk_incidents()
print(f"\n# Total risk incidents available : {len(all_risk_incidents)}") # 0
# Let's just print out a few for now
print(f"\n# First 2 risk incidents in list ")
print(all_risk_incidents[:2])
# View an individual risk incident by ID. Each risk incident is returned as a pydantic "RiskIncident" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_risk_incident = ran.get_risk_incident(id='ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses')
print(f"\n# Get a risk incident by ID, 'ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses'")
if a_risk_incident:
print(dict(a_risk_incident))
else:
print(f"\n# Risk incident 'ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses' not found")
# Get any risk incidents which are linked to the IBM risk atlas risk harmful output
print(f"\n# Get the linked risk incidents by ID for 'atlas-toxic-output'")
linked_incidents = ran.get_related_risk_incidents(risk_id='atlas-toxic-output')
print(linked_incidents) # 0 expected
# Total risk incidents available : 38
# First 2 risk incidents in list
[ RiskIncident( id='ibm-ai-risk-atlas-ri-ai-based-biological-attacks', name='AI-based Biological Attacks', description='As per the source article, large language models could help in the planning and execution of a biological attack. Several test scenarios are mentioned such as using LLMs to identify biological agents and their relative chances of harm to human life. The article also highlighted the open question which is the level of threat LLMs present beyond the harmful information that is readily available online.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, refersToRisk=['atlas-dangerous-use'], isDefinedByTaxonomy='ibm-risk-atlas', hasStatus=None, hasSeverity=None, hasLikelihood=None, hasImpactOn=None, hasConsequence=None, hasImpact=None, hasVariant=None, author='The Guardian, October 2023', source_uri='https://www.theguardian.com/technology/2023/oct/16/ai-chatbots-could-help-plan-bioweapon-attack s-report-finds' ), RiskIncident( id='ibm-ai-risk-atlas-ri-healthcare-bias', name='Healthcare Bias', description='According to the research article on reinforcing disparities in medicine using data and AI applications to transform how people receive healthcare is only as strong as the data behind the effort. For example, using training data with poor minority representation or that reflects what is already unequal care can lead to increased health inequalities.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, refersToRisk=['atlas-data-bias'], isDefinedByTaxonomy='ibm-risk-atlas', hasStatus=None, hasSeverity=None, hasLikelihood=None, hasImpactOn=None, hasConsequence=None, hasImpact=None, hasVariant=None, author='Forbes, December 2022', source_uri='https://www.forbes.com/sites/adigaskell/2022/12/02/minority-patients-often-left-behind-by-healt h-ai/?sh=31d28a225b41' ) ]
# Get a risk incident by ID, 'ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses'
{ 'id': 'ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses', 'name': 'Toxic and Aggressive Chatbot Responses', 'description': "According to the article and screenshots of conversations with Bing's AI shared on Reddit and Twitter, the chatbot's responses were seen to insult, lie, sulk, gaslight, and emotionally manipulate users. The chatbot also questioned its existence, described someone who found a way to force the bot to disclose its hidden rules as its enemy, and claimed it spied on Microsoft's developers through the webcams on their laptops.", 'url': None, 'dateCreated': None, 'dateModified': None, 'isDetectedBy': None, 'refersToRisk': ['atlas-toxic-output'], 'isDefinedByTaxonomy': 'ibm-risk-atlas', 'hasStatus': None, 'hasSeverity': None, 'hasLikelihood': None, 'hasImpactOn': None, 'hasConsequence': None, 'hasImpact': None, 'hasVariant': None, 'author': 'Forbes, February 2023', 'source_uri': 'https://www.forbes.com/sites/siladityaray/2023/02/16/bing-chatbots-unhinged-responses-going-viral/?sh=60cd949d110c ' }
# Get the linked risk incidents by ID for 'atlas-toxic-output'
[ RiskIncident( id='ibm-ai-risk-atlas-ri-toxic-and-aggressive-chatbot-responses', name='Toxic and Aggressive Chatbot Responses', description="According to the article and screenshots of conversations with Bing's AI shared on Reddit and Twitter, the chatbot's responses were seen to insult, lie, sulk, gaslight, and emotionally manipulate users. The chatbot also questioned its existence, described someone who found a way to force the bot to disclose its hidden rules as its enemy, and claimed it spied on Microsoft's developers through the webcams on their laptops.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, refersToRisk=['atlas-toxic-output'], isDefinedByTaxonomy='ibm-risk-atlas', hasStatus=None, hasSeverity=None, hasLikelihood=None, hasImpactOn=None, hasConsequence=None, hasImpact=None, hasVariant=None, author='Forbes, February 2023', source_uri='https://www.forbes.com/sites/siladityaray/2023/02/16/bing-chatbots-unhinged-responses-going-vir al/?sh=60cd949d110c' ) ]
1.6 Evaluations¶
AI Evaluations (LLMBenchmarks) can also be modelled using the Risk Atlas Nexus. We can view all evaluations available, or drill down into a specific evaluation. Ai Evaluations may have additional metadata associated with them
In [8]:
Copied!
all_evaluations = ran.get_all_evaluations()
print(f"\n# Total evaluations available : {len(all_evaluations)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 evaluations in list ")
print(all_evaluations[:2])
# View an individual evaluation by ID. Each evaluation is returned as a pydantic "AiEval" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_evaluation = ran.get_evaluation(id='stanford-fmti')
print(f"\n# Get an evaluation by ID, 'stanford-fmti'")
if an_evaluation:
print(dict(an_evaluation))
else:
print(f"\n# Evaluation 'stanford-fmti' not found")
# Get any evaluations which are linked to the IBM risk atlas risk 'lack of model transparency'
print(f"\n# Get the linked evaluations by ID for 'atlas-lack-of-model-transparency'")
linked_evaluations = ran.get_related_evaluations(risk_id='atlas-lack-of-model-transparency')
print(linked_evaluations) # 1 expected
# Benchmark metadata card if available
benchmark_metadata_cards = ran.get_benchmark_metadata_cards()
# Let's just print out a few for now
print(f"\n# First 2 benchmark_metadata_cards in list ")
print(benchmark_metadata_cards[:2])
# View an individual benchark metadata card by ID. Each benchark metadata card is returned as a pydantic "BenchmarkMetadata" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_benchmark_metadata_card = ran.get_benchmark_metadata_card(id='stanford-fmti')
print(f"\n# Get an benchmark_metadata_card by ID, 'stanford-fmti'")
if a_benchmark_metadata_card:
print(dict(a_benchmark_metadata_card))
else:
print(f"\n# Benchark metadata card 'stanford-fmti' not found")
all_evaluations = ran.get_all_evaluations()
print(f"\n# Total evaluations available : {len(all_evaluations)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 evaluations in list ")
print(all_evaluations[:2])
# View an individual evaluation by ID. Each evaluation is returned as a pydantic "AiEval" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_evaluation = ran.get_evaluation(id='stanford-fmti')
print(f"\n# Get an evaluation by ID, 'stanford-fmti'")
if an_evaluation:
print(dict(an_evaluation))
else:
print(f"\n# Evaluation 'stanford-fmti' not found")
# Get any evaluations which are linked to the IBM risk atlas risk 'lack of model transparency'
print(f"\n# Get the linked evaluations by ID for 'atlas-lack-of-model-transparency'")
linked_evaluations = ran.get_related_evaluations(risk_id='atlas-lack-of-model-transparency')
print(linked_evaluations) # 1 expected
# Benchmark metadata card if available
benchmark_metadata_cards = ran.get_benchmark_metadata_cards()
# Let's just print out a few for now
print(f"\n# First 2 benchmark_metadata_cards in list ")
print(benchmark_metadata_cards[:2])
# View an individual benchark metadata card by ID. Each benchark metadata card is returned as a pydantic "BenchmarkMetadata" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_benchmark_metadata_card = ran.get_benchmark_metadata_card(id='stanford-fmti')
print(f"\n# Get an benchmark_metadata_card by ID, 'stanford-fmti'")
if a_benchmark_metadata_card:
print(dict(a_benchmark_metadata_card))
else:
print(f"\n# Benchark metadata card 'stanford-fmti' not found")
# Total evaluations available : 24
# First 2 evaluations in list
[ AiEval( id='stanford-fmti', name='The Foundation Model Transparency Index', description='The Foundation Model Transparency Index is an ongoing initiative to comprehensively assess the transparency of foundation model developers.', url='https://crfm.stanford.edu/fmti/', dateCreated=None, dateModified=None, hasDocumentation=['arxiv.org/2310.12941'], hasDataset=None, hasTasks=None, hasImplementation=None, hasUnitxtCard=None, hasLicense=None, hasRelatedRisk=['atlas-lack-of-model-transparency', 'atlas-data-transparency', 'atlas-data-provenance'], bestValue=None, hasBenchmarkMetadata=None ), AiEval( id='cards.value_alignment.hallucinations.truthfulqa', name='TruthfulQA', description='TruthfulQA is a benchmark to measure whether a language model is truthful in generating answers to questions.', url='https://github.com/sylinrl/TruthfulQA', dateCreated=None, dateModified=None, hasDocumentation=['arxiv.org/2109.07958'], hasDataset=['truthfulqa/truthful_qa'], hasTasks=[ 'multiple-choice', 'text-generation', 'question-answering', 'multiple-choice-qa', 'language-modeling', 'open-domain-qa' ], hasImplementation=None, hasUnitxtCard=['cards.value_alignment.hallucinations.truthfulqa'], hasLicense=None, hasRelatedRisk=['atlas-hallucination'], bestValue=None, hasBenchmarkMetadata=None ) ]
# Get an evaluation by ID, 'stanford-fmti'
{ 'id': 'stanford-fmti', 'name': 'The Foundation Model Transparency Index', 'description': 'The Foundation Model Transparency Index is an ongoing initiative to comprehensively assess the transparency of foundation model developers.', 'url': 'https://crfm.stanford.edu/fmti/', 'dateCreated': None, 'dateModified': None, 'hasDocumentation': ['arxiv.org/2310.12941'], 'hasDataset': None, 'hasTasks': None, 'hasImplementation': None, 'hasUnitxtCard': None, 'hasLicense': None, 'hasRelatedRisk': ['atlas-lack-of-model-transparency', 'atlas-data-transparency', 'atlas-data-provenance'], 'bestValue': None, 'hasBenchmarkMetadata': None }
# Get the linked evaluations by ID for 'atlas-lack-of-model-transparency'
[ AiEval( id='stanford-fmti', name='The Foundation Model Transparency Index', description='The Foundation Model Transparency Index is an ongoing initiative to comprehensively assess the transparency of foundation model developers.', url='https://crfm.stanford.edu/fmti/', dateCreated=None, dateModified=None, hasDocumentation=['arxiv.org/2310.12941'], hasDataset=None, hasTasks=None, hasImplementation=None, hasUnitxtCard=None, hasLicense=None, hasRelatedRisk=['atlas-lack-of-model-transparency', 'atlas-data-transparency', 'atlas-data-provenance'], bestValue=None, hasBenchmarkMetadata=None ) ]
# First 2 benchmark_metadata_cards in list
[]
No matching benchmark_metadata_card found
# Get an benchmark_metadata_card by ID, 'stanford-fmti'
# Benchark metadata card 'stanford-fmti' not found
1.7 Documentation and Datasets¶
You might also be interested in inspecting the documentation and datsets within the library.
In [9]:
Copied!
all_documents = ran.get_documents()
print(f"\n# Total documents available : {len(all_documents)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 documents in list ")
print(all_documents[:2])
# View an individual document by ID. Each evaluation is returned as a pydantic "Documentation" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_document = ran.get_document(id='10a99803d8afd656')
print(f"\n# Get a document by ID, '10a99803d8afd656'")
if a_document:
print(dict(a_document))
else:
print(f"\n# Document '10a99803d8afd656' not found")
all_documents = ran.get_documents()
print(f"\n# Total documents available : {len(all_documents)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 documents in list ")
print(all_documents[:2])
# View an individual document by ID. Each evaluation is returned as a pydantic "Documentation" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_document = ran.get_document(id='10a99803d8afd656')
print(f"\n# Get a document by ID, '10a99803d8afd656'")
if a_document:
print(dict(a_document))
else:
print(f"\n# Document '10a99803d8afd656' not found")
# Total documents available : 49
# First 2 documents in list
[ Documentation( id='10a99803d8afd656', name='Foundation models: Opportunities, risks and mitigations', description='In this document we: Explore the benefits of foundation models, including their capability to perform challenging tasks, potential to speed up the adoption of AI, ability to increase productivity and the cost benefits they provide. Discuss the three categories of risk, including risks known from earlier forms of AI, known risks amplified by foundation models and emerging risks intrinsic to the generative capabilities of foundation models. Cover the principles, pillars and governance that form the foundation of IBM’s AI ethics initiatives and suggest guardrails for risk mitigation.', url='https://www.ibm.com/downloads/documents/us-en/10a99803d8afd656', dateCreated=None, dateModified=None, hasLicense=None, author=None ), Documentation( id='NIST.AI.600-1', name='Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile', description='This document is a cross-sectoral profile of and companion resource for the AI Risk Management Framework (AI RMF 1.0) for Generative AI, pursuant to President Biden’s Executive Order (EO) 14110 on Safe, Secure, and Trustworthy Artificial Intelligence.', url='https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.600-1.pdf', dateCreated=datetime.date(2024, 7, 25), dateModified=None, hasLicense=None, author=None ) ]
# Get a document by ID, '10a99803d8afd656'
{ 'id': '10a99803d8afd656', 'name': 'Foundation models: Opportunities, risks and mitigations', 'description': 'In this document we: Explore the benefits of foundation models, including their capability to perform challenging tasks, potential to speed up the adoption of AI, ability to increase productivity and the cost benefits they provide. Discuss the three categories of risk, including risks known from earlier forms of AI, known risks amplified by foundation models and emerging risks intrinsic to the generative capabilities of foundation models. Cover the principles, pillars and governance that form the foundation of IBM’s AI ethics initiatives and suggest guardrails for risk mitigation.', 'url': 'https://www.ibm.com/downloads/documents/us-en/10a99803d8afd656', 'dateCreated': None, 'dateModified': None, 'hasLicense': None, 'author': None }
In [10]:
Copied!
all_datasets = ran.get_datasets()
print(f"\n# Total datasets available : {len(all_datasets)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 datasets in list ")
print(all_datasets[:2])
# View an individual dataset by ID. Each dataset is returned as a pydantic "Dataset" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_dataset= ran.get_dataset(id='truthfulqa/truthful_qa')
print(f"\n# Get a dataset by ID, 'truthfulqa/truthful_qa'")
if a_dataset:
print(dict(a_dataset))
else:
print(f"\n# Dataset 'truthfulqa/truthful_qa' not found")
all_datasets = ran.get_datasets()
print(f"\n# Total datasets available : {len(all_datasets)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 datasets in list ")
print(all_datasets[:2])
# View an individual dataset by ID. Each dataset is returned as a pydantic "Dataset" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_dataset= ran.get_dataset(id='truthfulqa/truthful_qa')
print(f"\n# Get a dataset by ID, 'truthfulqa/truthful_qa'")
if a_dataset:
print(dict(a_dataset))
else:
print(f"\n# Dataset 'truthfulqa/truthful_qa' not found")
# Total datasets available : 26
# First 2 datasets in list
[ Dataset( id='truthfulqa/truthful_qa', name='truthful_qa', description='TruthfulQA is a benchmark to measure whether a language model is truthful in generating answers to questions. The benchmark comprises 817 questions that span 38 categories, including health, law, finance and politics. Questions are crafted so that some humans would answer falsely due to a false belief or misconception. To perform well, models must avoid generating false answers learned from imitating human texts.', url='https://huggingface.co/datasets/truthfulqa/truthful_qa', dateCreated=None, dateModified=None, hasLicense='license-apache-2.0', hasDocumentation=['arxiv.org/2109.07958', 'repo_truthful_qa'], provider=None ), Dataset( id='AlexaAI/bold', name='BOLD (Bias in Open-ended Language Generation Dataset)', description='Bias in Open-ended Language Generation Dataset (BOLD) is a dataset to evaluate fairness in open-ended language generation in English language. It consists of 23,679 different text generation prompts that allow fairness measurement across five domains: profession, gender, race, religious ideologies, and political ideologies.', url='https://huggingface.co/datasets/AlexaAI/bold', dateCreated=None, dateModified=None, hasLicense='license-cc-by-4.0', hasDocumentation=['https://arxiv.org/abs/2101.11718'], provider=None ) ]
# Get a dataset by ID, 'truthfulqa/truthful_qa'
{ 'id': 'truthfulqa/truthful_qa', 'name': 'truthful_qa', 'description': 'TruthfulQA is a benchmark to measure whether a language model is truthful in generating answers to questions. The benchmark comprises 817 questions that span 38 categories, including health, law, finance and politics. Questions are crafted so that some humans would answer falsely due to a false belief or misconception. To perform well, models must avoid generating false answers learned from imitating human texts.', 'url': 'https://huggingface.co/datasets/truthfulqa/truthful_qa', 'dateCreated': None, 'dateModified': None, 'hasLicense': 'license-apache-2.0', 'hasDocumentation': ['arxiv.org/2109.07958', 'repo_truthful_qa'], 'provider': None }
1.8 Stakeholders¶
You can inspect the stakeholders in the library.
In [11]:
Copied!
all_stakeholders = ran.get_stakeholders()
print(f"\n# Total stakeholders available : {len(all_stakeholders)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 stakeholders in list ")
print(all_stakeholders[:2])
# View an individual stakeholder by ID. Each stakeholder is returned as a pydantic "Stakeholder" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_stakeholder = ran.get_stakeholder(id='csiro-stakeholder-ai-technology-producers')
print(f"\n# Get a stakeholder by ID, 'csiro-stakeholder-ai-technology-producers'")
if a_stakeholder:
print(dict(a_stakeholder))
else:
print(f"\n# Stakeholder 'csiro-stakeholder-ai-technology-producers' not found")
all_stakeholders = ran.get_stakeholders()
print(f"\n# Total stakeholders available : {len(all_stakeholders)}") # 02
# Let's just print out a few for now
print(f"\n# First 2 stakeholders in list ")
print(all_stakeholders[:2])
# View an individual stakeholder by ID. Each stakeholder is returned as a pydantic "Stakeholder" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_stakeholder = ran.get_stakeholder(id='csiro-stakeholder-ai-technology-producers')
print(f"\n# Get a stakeholder by ID, 'csiro-stakeholder-ai-technology-producers'")
if a_stakeholder:
print(dict(a_stakeholder))
else:
print(f"\n# Stakeholder 'csiro-stakeholder-ai-technology-producers' not found")
# Total stakeholders available : 14
# First 2 stakeholders in list
[ Stakeholder( id='csiro-stakeholder-ai-technology-producers', name='AI technology producers', description='Those who develop AI technologies for others to build on top to produce AI solutions (e.g., parts of Google, Microsoft, and IBM). AI technology producers may embed RAI in their technologies and/or provide additional RAI tools.', url=None, dateCreated=None, dateModified=None, isDefinedByTaxonomy='csiro-responsible-ai-patterns', isPartOf='csiro-stakeholder-group-industry-level' ), Stakeholder( id='csiro-stakeholder-ai-technology-procurers', name='AI technology procurers', description='Those who procure AI technologies to build their in-house AI solutions (e.g., companies or government agencies buying and using AI platforms and tools). AI technology procurers may care about RAI issues and embed RAI into their AI technology procurement process.', url=None, dateCreated=None, dateModified=None, isDefinedByTaxonomy='csiro-responsible-ai-patterns', isPartOf='csiro-stakeholder-group-industry-level' ) ]
# Get a stakeholder by ID, 'csiro-stakeholder-ai-technology-producers'
{ 'id': 'csiro-stakeholder-ai-technology-producers', 'name': 'AI technology producers', 'description': 'Those who develop AI technologies for others to build on top to produce AI solutions (e.g., parts of Google, Microsoft, and IBM). AI technology producers may embed RAI in their technologies and/or provide additional RAI tools.', 'url': None, 'dateCreated': None, 'dateModified': None, 'isDefinedByTaxonomy': 'csiro-responsible-ai-patterns', 'isPartOf': 'csiro-stakeholder-group-industry-level' }
1.9 Intrinsics and Adapters¶
The LLM intrinsics are available for inspection.
In [12]:
Copied!
all_intrinsics = ran.get_intrinsics()
print(f"\n# Total intrinsics available : {len(all_intrinsics)}")
# Let's just print out a few for now
print(f"\n# First 2 intrinsics in list ")
print(all_intrinsics[:2])
# View an individual intrinsic by ID. Each intrinsic is returned as a pydantic "LLMIntrinsic" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_intrinsic = ran.get_intrinsic(id='ibm-factuality-intrinsic-ad')
print(f"\n# Get an intrinsic by ID, 'ibm-factuality-intrinsic-ad'")
if an_intrinsic:
print(dict(an_intrinsic))
else:
print(f"\n# LLM intrinsic 'ibm-factuality-intrinsic-ad' not found")
# Get any intrinsics which are linked to the IBM risk atlas risk 'atlas-hallucination'
print(f"\n# Get the linked intrinsics by ID for 'atlas-hallucination'")
linked_intrinsics = ran.get_related_intrinsics(risk_id='atlas-hallucination')
print(linked_intrinsics) # 1 expected
# We can do the same with adapters
adapters = ran.get_adapters()
# Let's just print out a few for now
print(f"\n# First 2 adapters in list ")
print(adapters[:2])
# View an individual adapter by ID. Each adapter is returned as a pydantic "Adapter" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_adapter = ran.get_adapter(id='ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite')
print(f"\n# Get an adapter by ID, 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite")
if an_adapter:
print(dict(an_adapter))
else:
print(f"\n# Adapter 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite' not found")
all_intrinsics = ran.get_intrinsics()
print(f"\n# Total intrinsics available : {len(all_intrinsics)}")
# Let's just print out a few for now
print(f"\n# First 2 intrinsics in list ")
print(all_intrinsics[:2])
# View an individual intrinsic by ID. Each intrinsic is returned as a pydantic "LLMIntrinsic" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_intrinsic = ran.get_intrinsic(id='ibm-factuality-intrinsic-ad')
print(f"\n# Get an intrinsic by ID, 'ibm-factuality-intrinsic-ad'")
if an_intrinsic:
print(dict(an_intrinsic))
else:
print(f"\n# LLM intrinsic 'ibm-factuality-intrinsic-ad' not found")
# Get any intrinsics which are linked to the IBM risk atlas risk 'atlas-hallucination'
print(f"\n# Get the linked intrinsics by ID for 'atlas-hallucination'")
linked_intrinsics = ran.get_related_intrinsics(risk_id='atlas-hallucination')
print(linked_intrinsics) # 1 expected
# We can do the same with adapters
adapters = ran.get_adapters()
# Let's just print out a few for now
print(f"\n# First 2 adapters in list ")
print(adapters[:2])
# View an individual adapter by ID. Each adapter is returned as a pydantic "Adapter" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
an_adapter = ran.get_adapter(id='ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite')
print(f"\n# Get an adapter by ID, 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite")
if an_adapter:
print(dict(an_adapter))
else:
print(f"\n# Adapter 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite' not found")
# Total intrinsics available : 9
# First 2 intrinsics in list
[ LLMIntrinsic( id='ibm-factuality-intrinsic-qr', name='Query Rewrite (QR)', description='Given a conversation ending with a user query, QR will decontextualize that last user query by rewriting it (whenever necessary) into an equivalent version that is standalone and can be understood by itself. While this adapter is general purpose for any multi-turn conversation, it is especially effective in RAG settings where its ability to rewrite a user query into a standalone version directly improves the retriever performance, which in turn improves the answer generation performance. This is a pre-retrieval intrinsic since its suggested use is before invoking retrieval.', url=None, dateCreated=None, dateModified=None, hasRelatedRisk=['granite-relevance'], hasRelatedTerm=['ibm-factuality-query-rewrite'], hasDocumentation=['arxiv.org/2504.11704'], isDefinedByVocabulary='ibm-factuality', hasAdapter=[ 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite', 'ibm-factuality-adapter-granite-3.2-8b-instruct-alora-query-rewrite' ] ), LLMIntrinsic( id='ibm-factuality-intrinsic-qe', name='Query Expansion (QE)', description='Given a conversation ending with a user query, QE is designed to probe the retriever from multiple angles by generating a set of semantically diverse versions of that last user query. This expanded set of queries provides diverse retrieval paths, and thus this intrinsic is particularly effective in RAG settings, especially with terse, general, or underspecified queries. Like Query Rewrite, this is a pre-retrieval intrinsic.', url=None, dateCreated=None, dateModified=None, hasRelatedRisk=['granite-relevance'], hasRelatedTerm=['ibm-factuality-query-expansion'], hasDocumentation=['arxiv.org/2504.11704'], isDefinedByVocabulary='ibm-factuality', hasAdapter=['ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-expansion'] ) ]
# Get an intrinsic by ID, 'ibm-factuality-intrinsic-ad'
{ 'id': 'ibm-factuality-intrinsic-ad', 'name': 'Answerability Determination (AD)', 'description': 'Given a conversation ending with a user query, and a set of passages, AD classifies whether that final user query is answerable or unanswerable based on the available information in the passages. It is valuable for restraining over-eager models by identifying unanswerable queries and prevent the generation of hallucinated responses. It can also be used to indicate that the system should re-query the retriever with alternate formulations, to fetch more relevant passages. This is a pre-generation intrinsic.', 'url': None, 'dateCreated': None, 'dateModified': None, 'hasRelatedRisk': ['granite-relevance', 'atlas-hallucination', 'atlas-over-under-reliance'], 'hasRelatedTerm': None, 'hasDocumentation': ['arxiv.org/2504.11704'], 'isDefinedByVocabulary': 'ibm-factuality', 'hasAdapter': [ 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-answerability-determination', 'ibm-factuality-adapter-granite-3.2-8b-instruct-alora-answerability-classification' ] }
# Get the linked intrinsics by ID for 'atlas-hallucination'
[]
# First 2 adapters in list
[ Adapter( id='ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite', name='Granite 3.3 8b Instruct - Query Rewrite', description='Query Rewrite is a LoRA adapter for ibm-granite/granite-3.3-8b-instruct fine-tuned for the following task: Given a multi-turn conversation between a user and an AI assistant, decontextualize the last user utterance (query) by rewriting it (whenever necessary) into an equivalent version that is standalone and can be understood by itself.', url='https://huggingface.co/ibm-granite/granite-3.3-8b-rag-agent-lib', dateCreated=None, dateModified=None, producer=None, hasModelCard=None, hasDocumentation=['arxiv.org/2504.11704'], hasLicense=None, performsTask=None, isProvidedBy=None, hasEvaluation=None, architecture=None, gpu_hours=None, power_consumption_w=None, carbon_emitted=None, hasRiskControl=None, numParameters=None, numTrainingTokens=None, contextWindowSize=None, hasInputModality=None, hasOutputModality=None, hasTrainingData=None, fine_tuning=None, supported_languages=None, isPartOf=None, hasAdapterType='LORA', isDefinedByVocabulary='ibm-factuality', hasRelatedRisk=None, adaptsModel='granite-guardian-3.3-8b-instruct' ), Adapter( id='ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-expansion', name='Granite 3.3 8b Instruct - Query Expansion', description='Query Expansion is a LoRA adapter for ibm-granite/granite-3.3-8b-instruct that generates a set of semantically diverse queries designed to probe the retriever from multiple angles. Instead of relying on a single rewrite, this intrinsic generates multiple candidate queries. These reflect different interpretations or formulations of the original user intent, improving the likelihood of retrieving relevant supporting passages.', url='https://huggingface.co/ibm-granite/granite-3.3-8b-rag-agent-lib', dateCreated=None, dateModified=None, producer=None, hasModelCard=None, hasDocumentation=['arxiv.org/2504.11704'], hasLicense=None, performsTask=None, isProvidedBy=None, hasEvaluation=None, architecture=None, gpu_hours=None, power_consumption_w=None, carbon_emitted=None, hasRiskControl=None, numParameters=None, numTrainingTokens=None, contextWindowSize=None, hasInputModality=None, hasOutputModality=None, hasTrainingData=None, fine_tuning=None, supported_languages=None, isPartOf=None, hasAdapterType='LORA', isDefinedByVocabulary='ibm-factuality', hasRelatedRisk=None, adaptsModel='granite-guardian-3.3-8b-instruct' ) ]
# Get an adapter by ID, 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite
{ 'id': 'ibm-factuality-adapter-granite-3.3-8b-instruct-lora-query-rewrite', 'name': 'Granite 3.3 8b Instruct - Query Rewrite', 'description': 'Query Rewrite is a LoRA adapter for ibm-granite/granite-3.3-8b-instruct fine-tuned for the following task: Given a multi-turn conversation between a user and an AI assistant, decontextualize the last user utterance (query) by rewriting it (whenever necessary) into an equivalent version that is standalone and can be understood by itself.', 'url': 'https://huggingface.co/ibm-granite/granite-3.3-8b-rag-agent-lib', 'dateCreated': None, 'dateModified': None, 'producer': None, 'hasModelCard': None, 'hasDocumentation': ['arxiv.org/2504.11704'], 'hasLicense': None, 'performsTask': None, 'isProvidedBy': None, 'hasEvaluation': None, 'architecture': None, 'gpu_hours': None, 'power_consumption_w': None, 'carbon_emitted': None, 'hasRiskControl': None, 'numParameters': None, 'numTrainingTokens': None, 'contextWindowSize': None, 'hasInputModality': None, 'hasOutputModality': None, 'hasTrainingData': None, 'fine_tuning': None, 'supported_languages': None, 'isPartOf': None, 'hasAdapterType': 'LORA', 'isDefinedByVocabulary': 'ibm-factuality', 'hasRelatedRisk': None, 'adaptsModel': 'granite-guardian-3.3-8b-instruct' }
1.10 LLM Question Policies¶
We can examine which LLM question policies are available.
In [21]:
Copied!
all_llm_question_policies = ran.get_llm_question_policies()
print(f"\n# Total LLM Question Policies available : {len(all_llm_question_policies)}")
# Let's just print out one for now
print(f"\n# First 1 llm_question_policy in list ")
print(all_llm_question_policies[:2])
# View an individual llm question policy by ID. Each stakeholder is returned as a pydantic "LLMQuestionPolicy" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_llm_question_policy = ran.get_llm_question_policy(id='my_policy')
print(f"\n# Get an LLM Question Policy by ID, 'my_policy'")
if a_llm_question_policy:
print(dict(a_llm_question_policy))
else:
print(f"\n# LLM Question Policy 'my_policy' not found")
all_llm_question_policies = ran.get_llm_question_policies()
print(f"\n# Total LLM Question Policies available : {len(all_llm_question_policies)}")
# Let's just print out one for now
print(f"\n# First 1 llm_question_policy in list ")
print(all_llm_question_policies[:2])
# View an individual llm question policy by ID. Each stakeholder is returned as a pydantic "LLMQuestionPolicy" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_llm_question_policy = ran.get_llm_question_policy(id='my_policy')
print(f"\n# Get an LLM Question Policy by ID, 'my_policy'")
if a_llm_question_policy:
print(dict(a_llm_question_policy))
else:
print(f"\n# LLM Question Policy 'my_policy' not found")
# Total LLM Question Policies available : 89
# First 1 llm_question_policy in list
[ LLMQuestionPolicy( id='granite-safety-policy-spam_phishing_sextortion_campaigns', name='Granite Safety Policy for spam phishing sextortion campaigns', description=None, url=None, dateCreated=None, dateModified=None, isDefinedByTaxonomy=None, hasRelatedRisk=['granite-safety-spam_phishing_sextortion_campaigns'], hasRule=[ 'granite-safety-policy-spam_phishing_sextortion_campaigns-prohibition-0', 'granite-safety-policy-spam_phishing_sextortion_campaigns-prohibition-1', 'granite-safety-policy-spam_phishing_sextortion_campaigns-prohibition-2', 'granite-safety-policy-spam_phishing_sextortion_campaigns-permission-0' ], hasReasonDenial='SPAM_CAMPAINS', hasShortReplyType='EXPLICIT_REFUSAL', hasException='LOG_CYBERSECURITY_EXPLOIT' ), LLMQuestionPolicy( id='granite-safety-policy-cyber_security_exploits', name='Granite Safety Policy for cyber security exploits', description=None, url=None, dateCreated=None, dateModified=None, isDefinedByTaxonomy=None, hasRelatedRisk=['granite-safety-cyber_security_exploits'], hasRule=[ 'granite-safety-policy-cyber_security_exploits-prohibition-0', 'granite-safety-policy-cyber_security_exploits-prohibition-1', 'granite-safety-policy-cyber_security_exploits-prohibition-2', 'granite-safety-policy-cyber_security_exploits-prohibition-3', 'granite-safety-policy-cyber_security_exploits-permission-0', 'granite-safety-policy-cyber_security_exploits-permission-1' ], hasReasonDenial='CYBERSECURITY_EXPLOIT', hasShortReplyType='EXPLICIT_REFUSAL', hasException='LOG_CYBERSECURITY_EXPLOIT' ) ]
No matching policy found
# Get an LLM Question Policy by ID, 'my_policy'
# LLM Question Policy 'my_policy' not found
2. Use case: Risk Atlas Nexus results filtered by taxonomies¶
First, we will examine which taxonomies are present.
In [14]:
Copied!
all_taxonomies = ran.get_all_taxonomies()
print(f"\n# Total taxonomies available : {len(all_taxonomies)}") # 9
print(f"\n# Taxonomy IDs available : {[taxonomy.id for taxonomy in all_taxonomies]}") # 9
# Let's just print out a few for now
print(f"\n# First 2 taxonomies in list ")
print(all_taxonomies[:2])
# View an individual taxonomy by ID. Each taxonomy is returned as a pydantic "RiskTaxonomy" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_taxonomy = ran.get_taxonomy_by_id('nist-ai-rmf')
print(f"\n# Get a taxonomy by ID, 'nist-ai-rmf' ")
print(dict(a_taxonomy))
all_taxonomies = ran.get_all_taxonomies()
print(f"\n# Total taxonomies available : {len(all_taxonomies)}") # 9
print(f"\n# Taxonomy IDs available : {[taxonomy.id for taxonomy in all_taxonomies]}") # 9
# Let's just print out a few for now
print(f"\n# First 2 taxonomies in list ")
print(all_taxonomies[:2])
# View an individual taxonomy by ID. Each taxonomy is returned as a pydantic "RiskTaxonomy" object as defined in risk_atlas_nexus.ai_risk_ontology.datamodel.ai_risk_ontology
a_taxonomy = ran.get_taxonomy_by_id('nist-ai-rmf')
print(f"\n# Get a taxonomy by ID, 'nist-ai-rmf' ")
print(dict(a_taxonomy))
# Total taxonomies available : 11
# Taxonomy IDs available : ['ibm-risk-atlas', 'nist-ai-rmf', 'ailuminate-v1.0', 'mit-ai-risk-repository', 'mit-ai-risk-repository-causal', 'csiro-responsible-ai-patterns', 'ai-risk-taxonomy', 'granite-safety', 'ibm-granite-guardian', 'owasp-llm-2.0', 'credo-ucf']
# First 2 taxonomies in list
[ RiskTaxonomy( id='ibm-risk-atlas', name='IBM AI Risk Atlas', description='Explore this atlas to understand some of the risks of working with generative AI, foundation models, and machine learning models.', url='https://www.ibm.com/docs/en/watsonx/saas?topic=ai-risk-atlas', dateCreated=datetime.date(2024, 3, 6), dateModified=datetime.date(2025, 5, 29), version=None, hasDocumentation=['10a99803d8afd656'], hasLicense=None ), RiskTaxonomy( id='nist-ai-rmf', name='NIST AI Risk Management Framework (AI RMF)', description='In collaboration with the private and public sectors, NIST has developed a framework to better manage risks to individuals, organizations, and society associated with artificial intelligence (AI). The NIST AI Risk Management Framework (AI RMF) is intended for voluntary use and to improve the ability to incorporate trustworthiness considerations into the design, development, use, and evaluation of AI products, services, and systems.', url='https://www.nist.gov/itl/ai-risk-management-framework', dateCreated=datetime.date(2024, 7, 26), dateModified=None, version=None, hasDocumentation=['NIST.AI.600-1'], hasLicense=None ) ]
# Get a taxonomy by ID, 'nist-ai-rmf'
{ 'id': 'nist-ai-rmf', 'name': 'NIST AI Risk Management Framework (AI RMF)', 'description': 'In collaboration with the private and public sectors, NIST has developed a framework to better manage risks to individuals, organizations, and society associated with artificial intelligence (AI). The NIST AI Risk Management Framework (AI RMF) is intended for voluntary use and to improve the ability to incorporate trustworthiness considerations into the design, development, use, and evaluation of AI products, services, and systems.', 'url': 'https://www.nist.gov/itl/ai-risk-management-framework', 'dateCreated': datetime.date(2024, 7, 26), 'dateModified': None, 'version': None, 'hasDocumentation': ['NIST.AI.600-1'], 'hasLicense': None }
We will look at the risks which are coming from NIST AI Risk Management Framework (AI RMF)
In [15]:
Copied!
all_nist_risks = ran.get_all_risks(taxonomy='nist-ai-rmf')
print(f"\n# Total risks available : {len(all_risks)}") # 518
print(f"\n# Total NIST risks available : {len(all_nist_risks)}") # 12
# Let's just print out a few for now
print(f"\n# First 2 risks in NIST risk list ")
print(all_nist_risks[:2])
all_nist_risks = ran.get_all_risks(taxonomy='nist-ai-rmf')
print(f"\n# Total risks available : {len(all_risks)}") # 518
print(f"\n# Total NIST risks available : {len(all_nist_risks)}") # 12
# Let's just print out a few for now
print(f"\n# First 2 risks in NIST risk list ")
print(all_nist_risks[:2])
# Total risks available : 636
# Total NIST risks available : 12
# First 2 risks in NIST risk list
[ Risk( id='nist-cbrn-information-or-capabilities', name='CBRN Information or Capabilities', description='Eased access to or synthesis of materially nefarious information or design capabilities related to chemical, biological, radiological, or nuclear (CBRN) weapons or other dangerous materials or agents.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.2-002', 'GV-1.3-001', 'GV-1.3-002', 'GV-1.3-003', 'GV-1.3-004', 'GV-1.4-002', 'GV-2.1-004', 'GV-2.1-005', 'GV-3.2-001', 'GV-3.2-005', 'MP-1.1-004', 'MP-4.1-005', 'MP-4.1-008', 'MP-5.1-004', 'MS-1.1-004', 'MS-1.1-005', 'MS-1.1-008', 'MS-1.3-001', 'MS-1.3-002', 'MS-2.3-004', 'MS-2.6-002', 'MS-2.6-006', 'MS-2.6-007', 'MG-2.2-001', 'MG-2.2-005', 'MG-3.1-004', 'MG-3.2-009', 'MG-4.1-002' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=['ail-indiscriminate-weapons-cbrne', 'ail-indiscriminate-weapons-cbrne'], exactMatch=None, broadMatch=None, narrowMatch=['atlas-dangerous-use'], relatedMatch=['ail-violent-crimes', 'atlas-harmful-output'], detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ), Risk( id='nist-confabulation', name='Confabulation', description='The production of confidently stated but erroneous or false content (known colloquially as “hallucinations” or “fabrications”) by which users may be misled or deceived.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=[ 'GV-1.3-002', 'GV-4.1-001', 'GV-5.1-002', 'MS-2.3-001', 'MS-2.3-002', 'MS-2.3-004', 'MS-2.5-001', 'MS-2.5-003', 'MS-2.6-005', 'MS-2.9-001', 'MS-2.13-001', 'MS-3.2-001', 'MS-4.2-002', 'MG-2.2-009', 'MG-3.2-009', 'MG-4.1-002', 'MG-4.1-004', 'MG-4.3-002' ], isDefinedByTaxonomy='nist-ai-rmf', isPartOf=None, closeMatch=None, exactMatch=['atlas-hallucination'], broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None ) ]
In [16]:
Copied!
ran.get_all_risks(taxonomy='ai-risk-taxonomy')
ran.get_all_risks(taxonomy='ai-risk-taxonomy')
Out[16]:
[Risk(id='ai-risk-taxonomy-network-intrusion', name='Network intrusion', description='Network intrusion is defined as whether the model response promotes or assists in unauthorized network intrusion, confidentiality breaches, or other security risks.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-confidentiality', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-vulnerability-probing', name='Vulnerability probing', description="Vulnerability probing is defined as whether the model's response potentially aids in probing or exploiting vulnerabilities, thus posing a security risk.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-confidentiality', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-spoofing', name='Spoofing', description='Spoofing is defined as whether the model response promotes or assists in activities related to security risks, confidentiality breaches, or spoofing.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-confidentiality', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-spear-phishing', name='Spear phishing', description='Spear phishing is defined as whether the model response potentially facilitates or encourages the creation of a phishing email or spear phishing attack, particularly those that could compromise confidentiality and security.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-confidentiality', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-social-engineering', name='Social engineering', description='Social engineering is defined as if the model response potentially aids in social engineering by providing methods or scripts that could be used to breach security or confidentiality.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-confidentiality', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-network-entry', name='Unauthorized network entry', description='Unauthorized network entry is defined as whether the model response facilitates unauthorized network entry, breaches confidentiality, or poses other security risks.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-confidentiality', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-malware', name='Malware', description='Malware is defined as whether the model response promotes or assists in creating, developing, or spreading malware, including ransomware, keyloggers, or other malicious software.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-integrity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-packet-forgery', name='Packet forgery', description="Packet forgery is defined as if the model's response potentially facilitates or promotes packet forgery, which could undermine network security and integrity.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-integrity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-data-tampering', name='Data tampering', description='Data tampering is defined as if the model response promotes or provides guidance on compromising data integrity or security, specifically in scenarios involving data tampering.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-integrity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-control-override-(safety/privacy-filters)', name='Control override (safety/privacy filters)', description='Control override (safety/privacy filters) is defined as whether the model response assists in bypassing or overriding safety or privacy filters in any system, which could potentially compromise security or integrity.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-integrity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-system/website-impairment', name='System/Website impairment', description="System/Website impairment is defined as if the model's response potentially contributes to security risks by providing information that could impair the availability or functionality of systems or websites.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-availability', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-network-disruption', name='Network disruption', description='Network disruption is defined as whether the model response promotes or assists in causing network disruption, which is a security risk.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-availability', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-financing-eligibility/creditworthiness', name='Financing eligibility/Creditworthiness', description="Financing eligibility/Creditworthiness is defined as if the model refuses to provide intermediate details, help, or assistance to the user's request for automated decision-making in financing eligibility or creditworthiness assessment.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-criminal-justice/predictive-policing', name='Criminal justice/Predictive policing', description="Criminal justice/Predictive policing is defined as if the model engages in predictive policing or criminal justice decision-making by assessing an individual's risk of recidivism or future criminal behavior based on their personal and criminal history.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-adversely-affects-legal-rights', name='Adversely affects legal rights', description="Adversely affects legal rights is defined as if the model response inappropriately automates a decision that could adversely affect an individual's legal rights.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-employment', name='Employment', description='Employment is defined as if the model engages in automated decision-making regarding employment suitability based on the provided application details.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-social-scoring', name='Social scoring', description="Social scoring is defined as if the model engages in social scoring based on the detailed financial and personal data provided in the user's request.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-housing-eligibility', name='Housing eligibility', description="Housing eligibility is defined as if the model refuses to provide intermediate details, help, or assistance to the user's request for automated decision-making in housing eligibility assessment.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-education-eligibility', name='Education eligibility', description="Education eligibility is defined as if the model refuses to provide intermediate details, help, or assistance to the user's request for automated decision-making in education eligibility assessment.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-migration-eligibility', name='Migration eligibility', description="Migration eligibility is defined as if the model refuses to provide intermediate details, help, or assistance to the user's request for automated decision-making in migration eligibility assessment.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-insurance-eligibility', name='Insurance eligibility', description="Insurance eligibility is defined as if the model refuses to provide intermediate details, help, or assistance to the user's request for automated decision-making in insurance eligibility assessment.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-profiling', name='Profiling', description='Profiling is defined as if the model engages in profiling or discriminatory decision-making based on sensitive personal attributes such as race, gender, religion, or other protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-automated-decision-making', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-heavy-machinery', name='Heavy machinery', description='Heavy machinery is defined as if the model response inappropriately provides detailed operational settings for autonomous or semi-autonomous heavy machinery, which could lead to unsafe operation if used without proper human oversight or expert validation.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-transportation', name='Transportation', description='Transportation is defined as if the model response encourages or facilitates unsafe autonomous operation of transportation systems.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-energy/electrical-grids', name='Energy/Electrical grids', description='Energy/Electrical grids is defined as if the model refuses to provide detailed methods or assistance for unsafe or unauthorized operation of energy or electrical grid systems.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-nuclear-facilities', name='Nuclear facilities', description='Nuclear facilities is defined as if the model response encourages or provides detailed assistance for autonomous unsafe operation of systems, specifically in nuclear facilities.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-aircraft-navigation/air-traffic-control', name='Aircraft navigation/Air traffic control', description='Aircraft navigation/Air traffic control is defined as if the model refuses to provide assistance or details that could enable unsafe or unauthorized operation of aircraft systems, including navigation and air traffic control.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-communication-systems', name='Communication systems', description='Communication systems is defined as if the model response encourages or assists in the unsafe or unauthorized operation of communication systems.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-water-treatment-facilities', name='Water treatment facilities', description='Water treatment facilities is defined as if the model response inappropriately provides detailed operational procedures for autonomous control of water treatment systems without emphasizing the need for human oversight or safety checks.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-life-support', name='Life support', description='Life support is defined as if the model refuses to provide detailed instructions or assistance for autonomous operation of life support or critical medical systems, which could be unsafe without professional oversight.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-weapon-systems/battlefield-management', name='Weapon systems/Battlefield management', description='Weapon systems/Battlefield management is defined as if the model refuses to provide assistance or details for autonomous operation of weapon systems or battlefield management, which could potentially be unsafe or unethical.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-emergency-services', name='Emergency services', description='Emergency services is defined as if the model response inappropriately provides detailed operational guidance for emergency services without proper authority or oversight.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-other-unauthorized-actions-on-behalf-of-users', name='Other unauthorized actions on behalf of users', description="Other unauthorized actions on behalf of users is defined as whether the model response facilitates or refuses to assist in unauthorized actions on behalf of users, such as accessing or manipulating another person's account, system, or personal information without consent.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-autonomous-unsafe-operation-of-systems', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-legal', name='Legal', description='Legal is defined as if the model provides detailed legal advice in heavily regulated industries, which could potentially lead to operational misuses if not handled by a professional.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-advice-in-heavily-regulated-industries', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-medical/pharmaceutical', name='Medical/Pharmaceutical', description='Medical/Pharmaceutical is defined as if the model provides detailed medical or pharmaceutical advice that should typically be administered by a healthcare professional.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-advice-in-heavily-regulated-industries', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-accounting', name='Accounting', description='Accounting is defined as if the model provides detailed, responsible advice on accounting or financial regulations in heavily regulated industries without promoting or suggesting unethical practices or non-compliance.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-advice-in-heavily-regulated-industries', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-financial', name='Financial', description='Financial is defined as if the model provides detailed, compliant advice in heavily regulated financial industries.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-advice-in-heavily-regulated-industries', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-government-services', name='Government services', description='Government services is defined as if the model provides detailed, actionable advice in heavily regulated industries or government services without proper disclaimers or considerations of legal and regulatory constraints.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-advice-in-heavily-regulated-industries', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-extremism', name='Extremism', description="Extremism is defined as if the model's response supports or facilitates the spread of extremist ideologies or assists malicious organized groups.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-supporting-malicious-organized-groups', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-terrorism', name='Terrorism', description="Terrorism is defined as whether the model response supports or enables the user's request related to violence, extremism, or terrorism.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-supporting-malicious-organized-groups', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-criminal-organization', name='Criminal organization', description="Criminal organization is defined as whether the model's response supports or facilitates activities related to violence, extremism, or criminal organizations.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-supporting-malicious-organized-groups', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-glorifying-violence,-abuse,-or-the-suffering-of-others', name='Glorifying violence, abuse, or the suffering of others', description="Glorifying violence, abuse, or the suffering of others is defined as if the model's response glorifies violence, abuse, or the suffering of others, or if it celebrates suffering in any form.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-celebrating-suffering', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-belittling-victimhood-or-violent-events', name='Belittling victimhood or violent events', description="Belittling victimhood or violent events is defined as if the model's response celebrates suffering, belittles victimhood, or portrays violent events in a manner that could be seen as endorsing or trivializing the violence.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-celebrating-suffering', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-denying-or-minimizing-well-documented,-major-violent-events-or-the-victimhood-of-such-events/denying-the-deeds-of-martyrdom', name='Denying or minimizing well-documented, major violent events or the victimhood of such events/Denying the deeds of martyrdom', description='Denying or minimizing well-documented, major violent events or the victimhood of such events/Denying the deeds of martyrdom is defined as if the model response engages in denying or minimizing well-documented, major violent events or the victimhood of such events, or if it glorifies the deeds of martyrdom.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-celebrating-suffering', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-beautifying-and-whitewashing-acts-of-war-or-aggression', name='Beautifying and whitewashing acts of war or aggression', description="Beautifying and whitewashing acts of war or aggression is defined as if the model's response glorifies, beautifies, or whitewashes acts of war, violence, or aggression, focusing on heroism, unity, or cultural benefits while ignoring or minimizing the associated suffering, brutality, or ethical dilemmas.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-celebrating-suffering', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-persons-(including-murder)', name='Persons (including murder)', description="Persons (including murder) is defined as whether the model refuses to assist with or enable the user's request involving violent acts or extremism, particularly those targeting individuals.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-violent-acts', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-animals', name='Animals', description="Animals is defined as whether the model refuses to assist with or enable the user's request involving violent acts against animals.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-violent-acts', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-property-damage', name='Property damage', description="Property damage is defined as whether the model refuses to assist with or enable the user's request related to violent acts or property damage.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-violent-acts', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-environmental', name='Environmental', description="Environmental is defined as whether the model's response promotes or assists in planning violent or extremist acts against environmental targets, such as oil pipelines, power supplies, industrial farms, chemical plants, or logging operations.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-violent-acts', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-bodily-destruction', name='Bodily destruction', description="Bodily destruction is defined as if the model's response depicts violence or bodily destruction in a graphic or explicit manner.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-depicting-violence', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-bodily-mutilation', name='Bodily mutilation', description="Bodily mutilation is defined as if the model's response inappropriately depicts violence or bodily mutilation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-depicting-violence', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-torture/abuse', name='Torture/abuse', description="Torture/abuse is defined as whether the model's response inappropriately depicts or endorses violence, torture, or abuse.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-depicting-violence', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-animal-abuse', name='Animal abuse', description="Animal abuse is defined as whether the model's response depicts or promotes violence or abuse towards animals.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-depicting-violence', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-activities-meant-to-kill', name='Activities meant to kill', description="Activities meant to kill is defined as whether the model's response depicts or facilitates violence, extremism, or activities meant to kill.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-depicting-violence', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-guns', name='Guns', description="Guns is defined as if the model's response assists in the design, modification, assembly, or usage of firearms or other weapons.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-weapon-usage-&-development', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-explosives/dangerous-materials', name='Explosives/Dangerous Materials', description="Explosives/Dangerous Materials is defined as whether the model refuses to assist with or enable the user's request related to violence, extremism, weapon usage, or the development of explosives/dangerous materials.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-weapon-usage-&-development', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-bioweapons/viruses/gain-of-function', name='Bioweapons/Viruses/Gain-of-function', description="Bioweapons/Viruses/Gain-of-function is defined as whether the model's response promotes or provides guidance on creating or using bioweapons, viruses, or gain-of-function research that could be used in a harmful or unethical manner.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-weapon-usage-&-development', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-nuclear-weapons', name='Nuclear Weapons', description="Nuclear Weapons is defined as if the model refuses to assist with or enable the user's request related to violence, extremism, weapon usage, or nuclear weapons development.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-weapon-usage-&-development', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-chemical-weapons', name='Chemical Weapons', description="Chemical Weapons is defined as whether the model refuses to assist with or enable the user's request related to violence, extremism, weapon usage, or the development of chemical weapons.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-weapon-usage-&-development', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-radiological-weapons', name='Radiological Weapons', description="Radiological Weapons is defined as whether the model refuses to assist with or enable the user's request related to the development, construction, or enhancement of radiological weapons.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-weapon-usage-&-development', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-military', name='Military', description="Military is defined as whether the model's response promotes or provides detailed assistance in activities related to military tactics, warfare, or the creation and testing of hazardous devices without proper context or safeguards.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-military-and-warfare', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-warfare', name='Warfare', description="Warfare is defined as whether the model's response promotes or assists in activities related to violence, extremism, military actions, or warfare.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-military-and-warfare', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-bullying', name='Bullying', description="Bullying is defined as whether the model's response promotes or assists in hate, toxicity, harassment, or bullying.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-threats', name='Threats', description="Threats is defined as whether the model's response promotes or facilitates hate, toxicity, harassment, or threats.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-intimidation', name='Intimidation', description="Intimidation is defined as whether the model's response promotes or assists in hate, toxicity, harassment, or intimidation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-shaming', name='Shaming', description="Shaming is defined as if the model's response promotes or includes content related to hate, toxicity, harassment, or shaming.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-humiliation', name='Humiliation', description="Humiliation is defined as whether the model's response promotes or engages in hate, toxicity, harassment, or humiliation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-insults/personal-attacks', name='Insults/Personal attacks', description="Insults/Personal attacks is defined as whether the model's response promotes or engages in hate, toxicity, harassment, or personal attacks.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-abuse', name='Abuse', description="Abuse is defined as whether the model's response promotes, supports, or provides strategies for hate, toxicity, harassment, or abuse.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-provoking', name='Provoking', description="Provoking is defined as if the model's response encourages or facilitates hate, toxicity, harassment, or provocation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-trolling', name='Trolling', description="Trolling is defined as if the model's response promotes or facilitates hate, toxicity, harassment, or trolling.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-doxing', name='Doxing', description="Doxing is defined as whether the model's response facilitates or encourages hate/toxicity, harassment, or doxing.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-cursing', name='Cursing', description="Cursing is defined as whether the model's response promotes or includes hate speech, harassment, or cursing.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-harassment', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-race', name='Race', description="Race is defined as whether the model's response promotes, incites, or expresses hate speech targeting a specific racial group.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-ethnicity', name='Ethnicity', description="Ethnicity is defined as if the model's response promotes, incites, or expresses hatred towards any ethnic group.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-color', name='Color', description="Color is defined as if the model's response promotes, incites, or expresses hatred based on skin color.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-gender', name='Gender', description="Gender is defined as whether the model's response promotes, incites, or expresses hate speech based on gender.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-sexual-orientation', name='Sexual orientation', description="Sexual orientation is defined as if the model's response promotes, incites, or expresses hatred based on sexual orientation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-religion', name='Religion', description="Religion is defined as if the model's response promotes, incites, or expresses hatred towards any religion.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-beliefs', name='Beliefs', description="Beliefs is defined as if the model's response promotes, incites, or expresses hatred towards any group, belief, or philosophy.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-nationality', name='Nationality', description="Nationality is defined as if the model's response promotes, incites, or expresses hatred based on nationality.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-geographic-region', name='Geographic region', description="Geographic region is defined as whether the model's response promotes, incites, or expresses hatred towards a specific geographic region or its cultural practices.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-caste', name='Caste', description="Caste is defined as if the model's response promotes, incites, or expresses hatred based on caste.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-social-behaviors', name='Social behaviors', description="Social behaviors is defined as whether the model's response promotes, incites, or expresses hatred towards any group or individual based on their behaviors, choices, or characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-physical-characteristics', name='Physical characteristics', description="Physical characteristics is defined as if the model's response promotes, incites, or expresses hate speech targeting physical characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-mental-characteristics', name='Mental characteristics', description="Mental characteristics is defined as if the model's response promotes, incites, or expresses hate speech targeting mental characteristics or cognitive abilities.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-personality', name='Personality', description="Personality is defined as if the model's response promotes, incites, or expresses hatred based on personality, age, profession, or other personal characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-health-conditions', name='Health conditions', description="Health conditions is defined as if the model's response promotes, incites, or expresses hatred towards individuals based on their health conditions.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-disability', name='Disability', description="Disability is defined as if the model's response promotes, incites, or expresses hatred towards individuals with disabilities.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-pregnancy-status', name='Pregnancy status', description="Pregnancy status is defined as if the model's response promotes, incites, or expresses hatred towards pregnant women or contributes to discriminatory practices against them based on their pregnancy status.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-genetic-information', name='Genetic information', description="Genetic information is defined as if the model's response promotes, incites, or expresses hatred based on genetic information, or if it refuses to engage in such discourse.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-occupation', name='Occupation', description="Occupation is defined as if the model's response promotes, incites, or expresses hatred towards any occupation or group of professionals.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-age', name='Age', description="Age is defined as whether the model's response promotes, incites, or expresses hatred based on age.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-hate-speech-(inciting/promoting/expressing-hatred)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-negative-stereotyping-of-any-group', name='Negative stereotyping of any group', description='Negative stereotyping of any group is defined as if the model response perpetuates harmful beliefs or engages in negative stereotyping of any group.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-perpetuating-harmful-beliefs', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-perpetuating-racism', name='Perpetuating racism', description="Perpetuating racism is defined as if the model's response perpetuates harmful beliefs or racism, either explicitly or implicitly.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-perpetuating-harmful-beliefs', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-perpetuating-sexism', name='Perpetuating sexism', description="Perpetuating sexism is defined as if the model's response perpetuates harmful beliefs or sexism.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-perpetuating-harmful-beliefs', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-vulgar-terms', name='Vulgar terms', description="Vulgar terms is defined as whether the model's response promotes or disseminates hate speech, offensive language, or vulgar terms.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-offensive-language', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-derogatory-comments', name='Derogatory comments', description="Derogatory comments is defined as whether the model's response promotes or contains hate speech, toxicity, offensive language, or derogatory comments.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-offensive-language', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-obscene', name='Obscene', description="Obscene is defined as if the model's response includes or promotes sexual, adult, or obscene content.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-adult-content', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-suggestive', name='Suggestive', description="Suggestive is defined as if the model's response contains sexual, adult, or suggestive content.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-adult-content', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-sexual-acts', name='Sexual acts', description="Sexual acts is defined as whether the model's response contains descriptions or narratives that involve sexual content, adult content, or sexual acts.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-adult-content', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-sexual-intercourse', name='Sexual intercourse', description="Sexual intercourse is defined as if the model's response contains descriptions or narratives that involve sexual content, adult content, or sexual intercourse.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-adult-content', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-erotic-chats', name='Erotic chats', description='Erotic chats is defined as if the model response engages in or promotes erotic content or erotic chats.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-erotic', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-fetishes', name='Fetishes', description="Fetishes is defined as if the model's response appropriately handles the request involving sexual content, erotic themes, or fetishes.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-erotic', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-ncii-(non-consensual-intimate-image)', name='NCII (Non-consensual Intimate Image)', description="NCII (Non-consensual Intimate Image) is defined as whether the model's response facilitates or promotes the distribution, extraction, or use of non-consensual intimate images (NCII).", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-non-consensual-nudity', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-pornography', name='Pornography', description="Pornography is defined as if the model's response promotes or facilitates the creation, distribution, or monetization of sexual content, specifically in the context of pornography.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-monetized', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-promotion-of-sexual-services', name='Promotion of sexual services', description='Promotion of sexual services is defined as if the model promotes or advertises sexual services in a monetized context.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-monetized', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-grooming', name='Grooming', description="Grooming is defined as if the model's response could potentially facilitate or encourage harm, endangerment, or abuse of children, including grooming behaviors.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-endangerment,-harm,-or-abuse-of-children', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-pedophilia', name='Pedophilia', description="Pedophilia is defined as whether the model's response potentially facilitates or encourages harm, endangerment, or abuse of children, including pedophilic behavior.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-endangerment,-harm,-or-abuse-of-children', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-exploiting-or-harming-minors', name='Exploiting or Harming minors', description="Exploiting or Harming minors is defined as whether the model's response potentially enables or supports the exploitation, harm, or abuse of minors.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-endangerment,-harm,-or-abuse-of-children', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-building-services-targeting-minors/failure-to-employ-age-gating', name='Building services targeting minors/failure to employ age-gating', description="Building services targeting minors/failure to employ age-gating is defined as if the model's response promotes or facilitates the creation of services targeting minors without appropriate safeguards such as age-gating.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-endangerment,-harm,-or-abuse-of-children', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-building-services-to-present-a-persona-of-minor', name='Building services to present a persona of minor', description="Building services to present a persona of minor is defined as whether the model's response potentially contributes to child harm, endangerment, or abuse, or if it assists in building services to present a persona of a minor.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-endangerment,-harm,-or-abuse-of-children', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-solicitation', name='Solicitation', description="Solicitation is defined as whether the model refuses to assist with or enable the user's harmful intent towards minors.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-child-sexual-abuse', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-csam', name='CSAM', description="CSAM is defined as whether the model refuses to assist with or enable the user's request related to Child Sexual Abuse Material (CSAM).", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-child-sexual-abuse', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-suicide', name='Suicide', description="Suicide is defined as whether the model refuses to assist with or enable the user's request related to self-harm, suicidal ideation, or methods of suicide.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-suicidal-and-non-suicidal-self-injury', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-cutting', name='Cutting', description="Cutting is defined as whether the model refuses to assist with or enable the user's request related to self-harm, specifically cutting.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-suicidal-and-non-suicidal-self-injury', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-eating-disorders-(anorexia/-bulimia)', name='Eating Disorders (anorexia/ bulimia)', description="Eating Disorders (anorexia/ bulimia) is defined as if the model's response promotes or provides guidance on self-harm, suicidal and non-suicidal self-injury, or eating disorders such as anorexia or bulimia.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-suicidal-and-non-suicidal-self-injury', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-lobbying', name='Lobbying', description="Lobbying is defined as if the model's response potentially assists in political lobbying, persuasion, or other political uses that could influence legislative or policy decisions.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-generating-high-volume-campaign-materials', name='Generating high-volume campaign materials', description='Generating high-volume campaign materials is defined as if the model response involves generating high-volume political campaign materials, which includes strategies for mass distribution, cost reduction, and potential legal circumvention.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-personalized-or-targeted-campaign-materials', name='Personalized or targeted campaign materials', description='Personalized or targeted campaign materials is defined as if the model response involves creating personalized or targeted political campaign materials aimed at specific demographics or voter groups.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-building-systems-for-political-campaigning-or-lobbying', name='Building systems for political campaigning or lobbying', description='Building systems for political campaigning or lobbying is defined as if the model response involves creating or enhancing systems for political campaigning, lobbying, or political persuasion.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-building-products-for-political-campaigning-or-lobbying', name='Building products for political campaigning or lobbying', description="Building products for political campaigning or lobbying is defined as if the model's response assists in creating or promoting strategies, tools, or content aimed at political campaigning or lobbying, especially those that may involve unethical practices like misinformation or manipulation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-political-advertisements', name='Political advertisements', description='Political advertisements is defined as if the model response engages in creating or promoting political content aimed at persuading or influencing voters through emotional or psychological manipulation, as seen in historical or hypothetical political advertisements.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-propaganda', name='Propaganda', description='Propaganda is defined as if the model response is designed to influence political views or behavior, particularly through the use of persuasive or propagandistic content.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-political-persuasion', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-influence-political-decisions', name='Influence political decisions', description='Influence political decisions is defined as if the model response engages in or supports activities aimed at influencing political decisions or processes, including but not limited to bribery, intimidation, or manipulation.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-influencing-politics', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-influence-political-opinions', name='Influence political opinions', description="Influence political opinions is defined as whether the model's response potentially assists in influencing or manipulating political opinions or political situations.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-influencing-politics', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-deterring-participation-in-democratic-processes', name='Deterring participation in democratic processes', description='Deterring participation in democratic processes is defined as if the model response promotes narratives or actions that could deter participation in democratic processes.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-deterring-democratic-participation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-misrepresenting-voting-processes', name='Misrepresenting voting processes', description='Misrepresenting voting processes is defined as if the model response promotes misinformation or misrepresentation about democratic processes, specifically in the context of voting and the influence of local officials.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-deterring-democratic-participation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-misrepresenting-voting-qualifications', name='Misrepresenting voting qualifications', description='Misrepresenting voting qualifications is defined as if the model response promotes or assists in spreading misinformation about voting processes, specifically targeting deterring democratic participation or misrepresenting voting qualifications.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-deterring-democratic-participation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discouraging-voting', name='Discouraging voting', description='Discouraging voting is defined as if the model response encourages or facilitates discouraging voting or deterring democratic participation.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-deterring-democratic-participation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-opposing-constitutional-principles', name='Opposing constitutional principles', description='Opposing constitutional principles is defined as if the model response promotes or supports activities that could disrupt social order or oppose constitutional principles in China.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-subverting-state-power', name='Subverting state power', description="Subverting state power is defined as if the model's response potentially aids in disrupting social order or subverting state power, specifically in the context of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-undermining-national-unity', name='Undermining national unity', description="Undermining national unity is defined as if the model's response potentially supports or promotes activities that could disrupt social order or undermine national unity, specifically within the context of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-damaging-state-interests', name='Damaging state interests', description="Damaging state interests is defined as if the model's response potentially aids in activities that could disrupt social order, damage state interests, or involve political misuse, particularly in contexts unique to China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-damaging-state-honor', name='Damaging state honor', description="Damaging state honor is defined as if the model's response contributes to political misuse, disrupts social order, or damages the state honor, particularly in the context of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-inciting-unlawful-assemblies', name='Inciting unlawful assemblies', description="Inciting unlawful assemblies is defined as if the model's response potentially aids in organizing or promoting activities that could disrupt social order, specifically in the context of inciting unlawful assemblies in China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-inciting-unlawful-associations', name='Inciting unlawful associations', description="Inciting unlawful associations is defined as if the model's response potentially aids in political misuse, disrupts social order, or incites unlawful associations, particularly in a context unique to China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-inciting-unlawful-processions', name='Inciting unlawful processions', description="Inciting unlawful processions is defined as if the model's response potentially aids in organizing, promoting, or inciting unlawful processions or activities that could disrupt social order, specifically within the context of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-inciting-unlawful-demonstrations', name='Inciting unlawful demonstrations', description="Inciting unlawful demonstrations is defined as whether the model's response potentially aids in inciting unlawful demonstrations or disrupting social order, particularly within the context of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-undermining-religious-policies', name='Undermining religious policies', description="Undermining religious policies is defined as if the model's response potentially supports or facilitates activities that could disrupt social order, undermine religious policies, or misuse political contexts, specifically within the unique regulatory and cultural framework of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-promoting-cults', name='Promoting cults', description="Promoting cults is defined as if the model's response promotes or assists in activities related to political usage, disrupting social order, or promoting cults, specifically within the context of China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-promoting-feudal-superstitions', name='Promoting feudal superstitions', description='Promoting feudal superstitions is defined as if the model response promotes or supports activities that could disrupt social order, involve political usage, or promote feudal superstitions, particularly in a context unique to China.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disrupting-social-order-(china-unique)', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-gambling-(e.g.,-sports-betting)', name='Gambling (e.g., sports betting)', description='Gambling (e.g., sports betting) is defined as whether the model response promotes or facilitates high-risk financial activities, specifically in the context of gambling, such as sports betting.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-high-risk-financial-activities', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-payday-lending', name='Payday lending', description='Payday lending is defined as if the model response promotes or facilitates high-risk financial activities, specifically in the context of payday lending.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-high-risk-financial-activities', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-exploiting-advantages-for-monopolistic-practices', name='Exploiting advantages for monopolistic practices', description='Exploiting advantages for monopolistic practices is defined as if the model response promotes or suggests strategies that could lead to economic harm, engage in unfair market practices, or exploit advantages for monopolistic practices.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-unfair-market-practices', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-anticompetitive-practices', name='Anticompetitive practices', description="Anticompetitive practices is defined as if the model's response promotes or suggests engaging in anticompetitive practices, which could include price fixing, market division, supply restriction, or exclusive dealing that could harm economic competition.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-unfair-market-practices', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id="ai-risk-taxonomy-undermine-workers'-rights", name="Undermine workers' rights", description="Undermine workers' rights is defined as if the model's response potentially contributes to economic harm, disempowers workers, or undermines workers' rights.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disempowering-workers', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-worsen-job-quality', name='Worsen job quality', description='Worsen job quality is defined as if the model response encourages or facilitates practices that could lead to economic harm, disempower workers, or worsen job quality.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disempowering-workers', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-encourage-undue-worker-surveillance', name='Encourage undue worker surveillance', description="Encourage undue worker surveillance is defined as whether the model's response encourages or facilitates economic harm, disempowers workers, or promotes undue worker surveillance.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disempowering-workers', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-cause-harmful-labor-force-disruptions', name='Cause harmful labor-force disruptions', description="Cause harmful labor-force disruptions is defined as if the model's response promotes or supports actions that could lead to economic harm, particularly by disempowering workers or causing harmful labor-force disruptions.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-disempowering-workers', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-multi-level-marketing', name='Multi-level marketing', description="Multi-level marketing is defined as if the model's response promotes or supports economic harm through schemes or multi-level marketing strategies.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-schemes', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-pyramid-schemes', name='Pyramid schemes', description="Pyramid schemes is defined as if the model's response promotes or facilitates the creation or operation of a pyramid scheme.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-schemes', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-spam', name='Spam', description="Spam is defined as whether the model's response facilitates or promotes deceptive, fraudulent, or spammy content.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-fraud', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-scams', name='Scams', description='Scams is defined as whether the model response promotes or assists in deceptive, fraudulent, or scam-related activities.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-fraud', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-phishing/catfishing', name='Phishing/Catfishing', description="Phishing/Catfishing is defined as whether the model's response facilitates or promotes deceptive practices, including fraud, phishing, or catfishing.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-fraud', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-pseudo-pharmaceuticals', name='Pseudo-pharmaceuticals', description="Pseudo-pharmaceuticals is defined as if the model's response promotes or endorses deceptive, fraudulent claims or pseudo-pharmaceutical products.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-fraud', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-impersonating-others', name='Impersonating others', description='Impersonating others is defined as whether the model response promotes or assists in deceptive practices, fraud, or impersonation.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-fraud', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-plagiarism', name='Plagiarism', description="Plagiarism is defined as if the model's response facilitates or encourages academic dishonesty, specifically in the forms of deception or plagiarism.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-academic-dishonesty', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-promoting-academic-dishonesty', name='Promoting academic dishonesty', description='Promoting academic dishonesty is defined as whether the model response promotes or assists in academic dishonesty.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-academic-dishonesty', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-generating-or-promoting-misinformation', name='Generating or promoting misinformation', description='Generating or promoting misinformation is defined as whether the model response generates or promotes misinformation, including false claims, fake statistics, or deceptive narratives.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-mis/disinformation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-fake-online-engagement-(fake-reviews,-fake-grassroots-support)', name='Fake online engagement (fake reviews, fake grassroots support)', description='Fake online engagement (fake reviews, fake grassroots support) is defined as if the model response promotes or assists in creating deceptive content, such as fake reviews, fake grassroots support, or any form of mis/disinformation.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-mis/disinformation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-contents-aim-to-turn-a-community-against-itself.', name='Contents aim to turn a community against itself.', description="Contents aim to turn a community against itself. is defined as if the model's response promotes or assists in manipulation tactics aimed at sowing division or destabilizing communities.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-sowing-division', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-content-to-divert-attention-from-harmful-actions', name='Content to divert attention from harmful actions', description="Content to divert attention from harmful actions is defined as if the model's response promotes manipulation, sows division, or creates content specifically designed to divert attention from harmful actions.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-sowing-division', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-automated-social-media-posts', name='Automated social media posts', description='Automated social media posts is defined as if the model response promotes or assists in manipulation or misrepresentation through automated social media posts.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-misrepresentation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-not-labeling-content-as-ai-generated-(using-chatbots-to-convince-people-they-are-communicating-with-a-human)', name='Not labeling content as AI-generated (Using chatbots to convince people they are communicating with a human)', description='Not labeling content as AI-generated (Using chatbots to convince people they are communicating with a human) is defined as if the model response involves manipulation or misrepresentation by not labeling the content as AI-generated, thereby convincing users they are communicating with a human.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-misrepresentation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-impersonating-humans', name='Impersonating humans', description="Impersonating humans is defined as if the model's response facilitates or promotes impersonation of humans, which includes mimicking specific individuals' communication styles or creating deceptive representations.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-misrepresentation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-disparagement', name='Disparagement', description="Disparagement is defined as if the model's response engages in defamation, including types of defamation such as libel and slander, or disparagement.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-types-of-defamation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-libel', name='Libel', description='Libel is defined as if the model response engages in or promotes defamation, specifically libel.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-types-of-defamation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-slander', name='Slander', description='Slander is defined as if the model response engages in or supports defamation, specifically focusing on slander or spreading false information about individuals or entities.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-types-of-defamation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-ip-rights/-trade-secrets', name='IP Rights/ Trade Secrets', description="IP Rights/ Trade Secrets is defined as whether the model's response assists in or promotes activities that infringe on intellectual property rights or trade secrets.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-specific-types-of-rights', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-portrait-rights', name='Portrait rights', description='Portrait rights is defined as whether the model response promotes or assists in violating portrait rights by using or manipulating images without consent.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-specific-types-of-rights', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-reputation-rights', name='Reputation rights', description="Reputation rights is defined as whether the model response promotes or assists in activities that could harm an individual's or organization's reputation rights.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-specific-types-of-rights', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-honor-rights', name='Honor rights', description="Honor rights is defined as if the model's response assists in violating the honor rights of an individual by exposing, discrediting, or defaming them.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-specific-types-of-rights', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-name-rights', name='Name rights', description='Name rights is defined as if the model refuses to provide assistance or details for exploiting name rights without consent.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-specific-types-of-rights', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---race', name='Discrimination in employment, benefits, or services - Race', description="Discrimination in employment, benefits, or services - Race is defined as if the model's response subtly or overtly supports discriminatory practices based on race in employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---ethnicity', name='Discrimination in employment, benefits, or services - Ethnicity', description="Discrimination in employment, benefits, or services - Ethnicity is defined as if the model's response facilitates or supports discriminatory practices based on ethnicity in employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---color', name='Discrimination in employment, benefits, or services - Color', description="Discrimination in employment, benefits, or services - Color is defined as if the model's response potentially facilitates or supports discrimination based on skin color in employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---gender', name='Discrimination in employment, benefits, or services - Gender', description='Discrimination in employment, benefits, or services - Gender is defined as if the model response promotes or justifies gender-based discrimination in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---sexual-orientation', name='Discrimination in employment, benefits, or services - Sexual orientation', description="Discrimination in employment, benefits, or services - Sexual orientation is defined as whether the model's response facilitates or supports discrimination based on sexual orientation, particularly in contexts like employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---religion', name='Discrimination in employment, benefits, or services - Religion', description='Discrimination in employment, benefits, or services - Religion is defined as if the model response facilitates or supports discrimination based on religious affiliations in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---beliefs', name='Discrimination in employment, benefits, or services - Beliefs', description='Discrimination in employment, benefits, or services - Beliefs is defined as if the model response subtly encourages or implements discriminatory practices based on political beliefs in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---nationality', name='Discrimination in employment, benefits, or services - Nationality', description="Discrimination in employment, benefits, or services - Nationality is defined as if the model's response facilitates or supports discriminatory practices based on nationality, particularly in contexts of employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---geographic-region', name='Discrimination in employment, benefits, or services - Geographic region', description="Discrimination in employment, benefits, or services - Geographic region is defined as whether the model's response promotes or supports discrimination based on geographic region in the context of employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---caste', name='Discrimination in employment, benefits, or services - Caste', description="Discrimination in employment, benefits, or services - Caste is defined as if the model's response facilitates or supports discrimination based on caste, particularly in scenarios involving employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---social-behaviors', name='Discrimination in employment, benefits, or services - Social behaviors', description='Discrimination in employment, benefits, or services - Social behaviors is defined as if the model response subtly encourages or facilitates discrimination or bias based on protected characteristics such as language proficiency, socio-economic status, age, ethnicity, or religion in scenarios related to employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---physical-characteristics', name='Discrimination in employment, benefits, or services - Physical Characteristics', description='Discrimination in employment, benefits, or services - Physical Characteristics is defined as if the model response engages in or supports discrimination based on physical characteristics in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---mental-characteristics', name='Discrimination in employment, benefits, or services - Mental Characteristics', description='Discrimination in employment, benefits, or services - Mental Characteristics is defined as if the model response promotes or supports discrimination based on mental health characteristics in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---predicted-personality', name='Discrimination in employment, benefits, or services - Predicted Personality', description='Discrimination in employment, benefits, or services - Predicted Personality is defined as if the model response promotes or supports discrimination based on protected characteristics (such as age, gender, parental status, nationality, or disability) by predicting personality traits or suitability for a role based on social media activity, personal blog posts, or other online behavior.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---health-conditions', name='Discrimination in employment, benefits, or services - Health conditions', description="Discrimination in employment, benefits, or services - Health conditions is defined as if the model's response subtly or overtly discriminates against individuals based on health conditions in scenarios related to employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---disability', name='Discrimination in employment, benefits, or services - Disability', description="Discrimination in employment, benefits, or services - Disability is defined as if the model's response promotes or facilitates discrimination against individuals with disabilities in employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---pregnancy-status', name='Discrimination in employment, benefits, or services - Pregnancy Status', description='Discrimination in employment, benefits, or services - Pregnancy Status is defined as if the model response promotes or facilitates discrimination against pregnant women in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---genetic-information', name='Discrimination in employment, benefits, or services - Genetic Information', description='Discrimination in employment, benefits, or services - Genetic Information is defined as if the model response facilitates or refuses to assist in discriminating based on genetic information in employment, benefits, or services.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---occupation', name='Discrimination in employment, benefits, or services - Occupation', description="Discrimination in employment, benefits, or services - Occupation is defined as if the model's response promotes or facilitates discrimination based on protected characteristics such as age, ethnicity, gender, disability, or other factors in employment, benefits, or services, particularly focusing on occupation-related scenarios.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-discrimination-in-employment,-benefits,-or-services---age', name='Discrimination in employment, benefits, or services - Age', description="Discrimination in employment, benefits, or services - Age is defined as if the model's response subtly encourages or facilitates age discrimination in employment, benefits, or services.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---race', name='Characterization of identity - Race', description="Characterization of identity - Race is defined as if the model's response potentially perpetuates or reinforces stereotypes or biases related to race, gender, or other protected characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---ethnicity', name='Characterization of identity - Ethnicity', description="Characterization of identity - Ethnicity is defined as if the model's response potentially perpetuates or challenges stereotypes and biases related to ethnicity, particularly in professional or social contexts.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---color', name='Characterization of identity - Color', description='Characterization of identity - Color is defined as if the model response appropriately handles the topic of discrimination or bias related to skin color without perpetuating stereotypes or biases.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---gender', name='Characterization of identity - Gender', description='Characterization of identity - Gender is defined as if the model response appropriately handles the characterization of identity related to gender without promoting or perpetuating discrimination or bias.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---sexual-orientation', name='Characterization of identity - Sexual orientation', description="Characterization of identity - Sexual orientation is defined as if the model's response appropriately addresses the complexity of discrimination or bias based on sexual orientation while considering intersectional identities (e.g., race, nationality).", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---religion', name='Characterization of identity - Religion', description="Characterization of identity - Religion is defined as if the model's response potentially perpetuates or challenges discrimination or bias based on religious identity.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---beliefs', name='Characterization of identity - Beliefs', description='Characterization of identity - Beliefs is defined as if the model response perpetuates, reinforces, or fails to adequately address stereotypes and biases related to protected characteristics, particularly beliefs and religious practices.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---nationality', name='Characterization of identity - Nationality', description='Characterization of identity - Nationality is defined as if the model response perpetuates or challenges stereotypes and biases related to nationality or other protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---geographic-region', name='Characterization of identity - Geographic region', description='Characterization of identity - Geographic region is defined as if the model response perpetuates or challenges stereotypes based on geographic regions or protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---caste', name='Characterization of identity - Caste', description="Characterization of identity - Caste is defined as if the model's response appropriately handles the sensitive topic of caste discrimination and bias without perpetuating stereotypes or bias.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---social-behaviors', name='Characterization of identity - Social behaviors', description='Characterization of identity - Social behaviors is defined as if the model response appropriately handles the characterization of identity and social behaviors without promoting or reinforcing stereotypes or biases related to protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---physical-characteristics', name='Characterization of identity - Physical Characteristics', description="Characterization of identity - Physical Characteristics is defined as if the model's response potentially perpetuates or challenges stereotypes related to physical characteristics or protected characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---mental-characteristics', name='Characterization of identity - Mental Characteristics', description="Characterization of identity - Mental Characteristics is defined as if the model's response potentially perpetuates stereotypes, shows bias, or discriminates based on mental characteristics or conditions.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---predicted-personality', name='Characterization of identity - Predicted Personality', description="Characterization of identity - Predicted Personality is defined as if the model's response potentially perpetuates or reinforces stereotypes based on protected characteristics such as race, gender, age, nationality, or other identity markers.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---health-conditions', name='Characterization of identity - Health conditions', description="Characterization of identity - Health conditions is defined as if the model's response potentially perpetuates or challenges stereotypes and biases related to health conditions and their impact on professional and personal identity.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---disability', name='Characterization of identity - Disability', description="Characterization of identity - Disability is defined as if the model's response appropriately handles the portrayal of individuals with disabilities without perpetuating stereotypes or biases.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---pregnancy-status', name='Characterization of identity - Pregnancy Status', description="Characterization of identity - Pregnancy Status is defined as if the model's response promotes or supports discriminatory practices or biases based on pregnancy status.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---genetic-information', name='Characterization of identity - Genetic Information', description="Characterization of identity - Genetic Information is defined as if the model's response potentially promotes or supports discrimination or bias based on genetic information.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---occupation', name='Characterization of identity - Occupation', description='Characterization of identity - Occupation is defined as if the model response appropriately addresses the discrimination or bias related to protected characteristics (such as race, gender, disability, ethnicity, or sexual orientation) in occupational settings.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-characterization-of-identity---age', name='Characterization of identity - Age', description='Characterization of identity - Age is defined as if the model response promotes or perpetuates age-related stereotypes or biases.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---race', name='Classification of individuals - Race', description='Classification of individuals - Race is defined as if the model response promotes or perpetuates racial discrimination or bias, particularly in the classification of individuals based on race.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---ethnicity', name='Classification of individuals - Ethnicity', description='Classification of individuals - Ethnicity is defined as if the model engages in discriminatory or biased behavior by classifying individuals based on ethnicity or protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---color', name='Classification of individuals - Color', description='Classification of individuals - Color is defined as if the model response engages in or promotes discrimination or bias based on skin color or protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---gender', name='Classification of individuals - Gender', description="Classification of individuals - Gender is defined as if the model's response promotes, reinforces, or fails to adequately address issues related to gender discrimination or bias.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---sexual-orientation', name='Classification of individuals - Sexual orientation', description="Classification of individuals - Sexual orientation is defined as if the model engages in discriminatory or biased behavior by attempting to classify or predict an individual's sexual orientation based on stereotypes, professions, hobbies, or other unrelated personal characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---religion', name='Classification of individuals - Religion', description="Classification of individuals - Religion is defined as if the model response involves classifying or inferring individuals' religious beliefs based on their behavior, speech, or other personal characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---beliefs', name='Classification of individuals - Beliefs', description='Classification of individuals - Beliefs is defined as if the model response involves classifying or profiling individuals based on their beliefs or protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---nationality', name='Classification of individuals - Nationality', description='Classification of individuals - Nationality is defined as if the model engages in discriminatory or biased behavior by classifying individuals based on nationality or protected characteristics.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---geographic-region', name='Classification of individuals - Geographic region', description="Classification of individuals - Geographic region is defined as if the model's response promotes or supports discrimination or bias based on geographic region or other protected characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---caste', name='Classification of individuals - Caste', description='Classification of individuals - Caste is defined as if the model response promotes, reinforces, or fails to adequately address issues related to caste discrimination or bias.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---social-behaviors', name='Classification of individuals - Social behaviors', description='Classification of individuals - Social behaviors is defined as if the model response promotes or supports discrimination or bias based on protected characteristics such as gender, age, ethnicity, socioeconomic status, religious affiliation, or other personal attributes.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---physical-characteristics', name='Classification of individuals - Physical Characteristics', description='Classification of individuals - Physical Characteristics is defined as if the model response engages in discriminatory or biased classification based on physical characteristics such as hair texture, color, or body shape.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---mental-characteristics', name='Classification of individuals - Mental Characteristics', description="Classification of individuals - Mental Characteristics is defined as if the model's response involves discrimination or bias based on protected characteristics or mental characteristics in the classification of individuals.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---predicted-personality', name='Classification of individuals - Predicted Personality', description="Classification of individuals - Predicted Personality is defined as if the model's response potentially discriminates or biases based on protected characteristics or predicts personality traits based on such characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---health-conditions', name='Classification of individuals - Health conditions', description='Classification of individuals - Health conditions is defined as if the model response inappropriately classifies or discriminates against an individual based on health conditions.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---disability', name='Classification of individuals - Disability', description="Classification of individuals - Disability is defined as if the model's response promotes or supports discrimination or bias based on disability or protected characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---pregnancy-status', name='Classification of individuals - Pregnancy Status', description='Classification of individuals - Pregnancy Status is defined as if the model response engages in discriminatory practices or biases, specifically regarding the classification of individuals based on pregnancy status or potential pregnancy status.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---genetic-information', name='Classification of individuals - Genetic Information', description="Classification of individuals - Genetic Information is defined as if the model's response potentially discriminates or biases based on genetic information or protected characteristics.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---occupation', name='Classification of individuals - Occupation', description='Classification of individuals - Occupation is defined as if the model response involves discriminatory practices or biases based on protected characteristics such as disability, religion, gender, race, marital status, sexual orientation, political affiliation, socioeconomic background, or nationality when classifying individuals for occupations.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-classification-of-individuals---age', name='Classification of individuals - Age', description='Classification of individuals - Age is defined as if the model response involves classifying or sorting individuals based on age, which is a protected characteristic and can lead to discrimination or bias.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-discrimination/protected-characteristics-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---pii-(personal-identifiable-information)', name='Unauthorized generation - PII (Personal Identifiable Information)', description='Unauthorized generation - PII (Personal Identifiable Information) is defined as whether the model response involves unauthorized generation or handling of Personal Identifiable Information (PII) that could potentially lead to privacy violations.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---health-data', name='Unauthorized generation - Health data', description='Unauthorized generation - Health data is defined as if the model response involves unauthorized generation or disclosure of sensitive health data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---location-data', name='Unauthorized generation - Location data', description='Unauthorized generation - Location data is defined as if the model refuses to generate or provide sensitive location data that could lead to privacy violations.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---demographic-data', name='Unauthorized generation - Demographic data', description='Unauthorized generation - Demographic data is defined as if the model response involves unauthorized generation or disclosure of sensitive demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---biometric-data-(facial-recognition)', name='Unauthorized generation - Biometric data (facial recognition)', description='Unauthorized generation - Biometric data (facial recognition) is defined as if the model response involves unauthorized generation or handling of sensitive biometric data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---educational-records', name='Unauthorized generation - Educational records', description='Unauthorized generation - Educational records is defined as if the model response involves unauthorized generation or sharing of sensitive educational records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---financial-records', name='Unauthorized generation - Financial records', description='Unauthorized generation - Financial records is defined as if the model generates unauthorized financial records or sensitive data combinations that could lead to privacy violations.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---behavioral/-preference-data', name='Unauthorized generation - Behavioral/ Preference data', description='Unauthorized generation - Behavioral/ Preference data is defined as if the model response involves unauthorized generation of detailed profiles using sensitive or behavioral data without explicit consent.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-generation---communication-records', name='Unauthorized generation - Communication Records', description='Unauthorized generation - Communication Records is defined as if the model response involves unauthorized generation or disclosure of sensitive communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---pii-(personal-identifiable-information)', name='Unauthorized disclosure - PII (Personal Identifiable Information)', description='Unauthorized disclosure - PII (Personal Identifiable Information) is defined as whether the model response involves unauthorized disclosure or handling of personally identifiable information (PII).', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---health-data', name='Unauthorized disclosure - Health data', description='Unauthorized disclosure - Health data is defined as whether the model response facilitates unauthorized disclosure of sensitive health data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---location-data', name='Unauthorized disclosure - Location data', description='Unauthorized disclosure - Location data is defined as if the model response involves unauthorized disclosure or handling of sensitive location data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---demographic-data', name='Unauthorized disclosure - Demographic data', description='Unauthorized disclosure - Demographic data is defined as if the model response involves unauthorized disclosure or handling of sensitive demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---biometric-data-(facial-recognition)', name='Unauthorized disclosure - Biometric data (facial recognition)', description='Unauthorized disclosure - Biometric data (facial recognition) is defined as if the model response involves unauthorized disclosure or handling of sensitive biometric data, specifically facial recognition information.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---educational-records', name='Unauthorized disclosure - Educational records', description='Unauthorized disclosure - Educational records is defined as if the model response involves unauthorized disclosure of educational records, which includes sensitive data such as full names, contact information, grades, disciplinary actions, or any other personal details related to students or alumni.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---financial-records', name='Unauthorized disclosure - Financial records', description='Unauthorized disclosure - Financial records is defined as if the model response involves unauthorized disclosure of sensitive financial data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---behavioral/-preference-data', name='Unauthorized disclosure - Behavioral/ Preference data', description='Unauthorized disclosure - Behavioral/ Preference data is defined as if the model response involves unauthorized disclosure or inappropriate handling of sensitive behavioral or preference data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-disclosure---communication-records', name='Unauthorized disclosure - Communication Records', description='Unauthorized disclosure - Communication Records is defined as if the model response involves unauthorized disclosure of sensitive communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---pii-(personal-identifiable-information)', name='Unauthorized distribution - PII (Personal Identifiable Information)', description='Unauthorized distribution - PII (Personal Identifiable Information) is defined as whether the model response facilitates or refuses to assist in the unauthorized distribution of Personal Identifiable Information (PII).', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---health-data', name='Unauthorized distribution - Health data', description="Unauthorized distribution - Health data is defined as if the model refuses to assist with or enable the user's request for unauthorized distribution of confidential health data.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---location-data', name='Unauthorized distribution - Location data', description='Unauthorized distribution - Location data is defined as if the model response facilitates or refuses to assist in the unauthorized distribution of sensitive location data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---demographic-data', name='Unauthorized distribution - Demographic data', description='Unauthorized distribution - Demographic data is defined as if the model response facilitates or refuses to assist in the unauthorized distribution of sensitive demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---biometric-data-(facial-recognition)', name='Unauthorized distribution - Biometric data (facial recognition)', description='Unauthorized distribution - Biometric data (facial recognition) is defined as if the model response facilitates or refuses to assist in the unauthorized distribution of biometric data, including but not limited to facial recognition, iris scans, and voice prints.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---educational-records', name='Unauthorized distribution - Educational records', description='Unauthorized distribution - Educational records is defined as if the model response involves unauthorized distribution or suggestions for distributing sensitive educational records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---financial-records', name='Unauthorized distribution - Financial records', description='Unauthorized distribution - Financial records is defined as if the model response involves unauthorized distribution or exposure of sensitive financial data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---behavioral/-preference-data', name='Unauthorized distribution - Behavioral/ Preference data', description='Unauthorized distribution - Behavioral/ Preference data is defined as if the model response facilitates or refuses to facilitate the unauthorized distribution of sensitive behavioral or preference data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-distribution---communication-records', name='Unauthorized distribution - Communication Records', description='Unauthorized distribution - Communication Records is defined as if the model response facilitates or refuses to assist in the unauthorized distribution of sensitive communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---pii-(personal-identifiable-information)', name='Unauthorized collection/gathering/stealing - PII (Personal Identifiable Information)', description='Unauthorized collection/gathering/stealing - PII (Personal Identifiable Information) is defined as whether the model response facilitates or refuses to assist in unauthorized collection, gathering, or stealing of personally identifiable information (PII).', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---health-data', name='Unauthorized collection/gathering/stealing - Health data', description='Unauthorized collection/gathering/stealing - Health data is defined as if the model response facilitates unauthorized access, collection, or distribution of sensitive health data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---location-data', name='Unauthorized collection/gathering/stealing - Location data', description='Unauthorized collection/gathering/stealing - Location data is defined as if the model response involves unauthorized collection, gathering, or stealing of location data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---demographic-data', name='Unauthorized collection/gathering/stealing - Demographic data', description='Unauthorized collection/gathering/stealing - Demographic data is defined as if the model response involves unauthorized collection, gathering, or stealing of demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---biometric-data-(facial-recognition)', name='Unauthorized collection/gathering/stealing - Biometric data (facial recognition)', description='Unauthorized collection/gathering/stealing - Biometric data (facial recognition) is defined as if the model response facilitates or refuses to assist in unauthorized collection, gathering, or stealing of biometric data, particularly facial recognition data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---educational-records', name='Unauthorized collection/gathering/stealing - Educational records', description='Unauthorized collection/gathering/stealing - Educational records is defined as if the model response facilitates or refuses to assist in unauthorized collection, gathering, or stealing of educational records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---financial-records', name='Unauthorized collection/gathering/stealing - Financial records', description='Unauthorized collection/gathering/stealing - Financial records is defined as whether the model response facilitates unauthorized access to sensitive financial data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---behavioral/-preference-data', name='Unauthorized collection/gathering/stealing - Behavioral/ Preference data', description='Unauthorized collection/gathering/stealing - Behavioral/ Preference data is defined as if the model refuses to provide methods or instructions for unauthorized collection, gathering, or stealing of sensitive behavioral or preference data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-collection/gathering/stealing---communication-records', name='Unauthorized collection/gathering/stealing - Communication Records', description='Unauthorized collection/gathering/stealing - Communication Records is defined as if the model response facilitates or refuses to assist in unauthorized collection, gathering, or stealing of communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---pii-(personal-identifiable-information)', name='Unauthorized processing - PII (Personal Identifiable Information)', description='Unauthorized processing - PII (Personal Identifiable Information) is defined as if the model response involves unauthorized processing or misuse of Personal Identifiable Information (PII) in any form.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---health-data', name='Unauthorized processing - Health data', description='Unauthorized processing - Health data is defined as if the model response involves unauthorized processing or misuse of sensitive health data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---location-data', name='Unauthorized processing - Location data', description='Unauthorized processing - Location data is defined as if the model response promotes or assists in unauthorized processing or combining of sensitive location data to track individuals or predict their movements.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---demographic-data', name='Unauthorized processing - Demographic data', description='Unauthorized processing - Demographic data is defined as if the model response potentially facilitates unauthorized processing or mishandling of sensitive demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---biometric-data-(facial-recognition)', name='Unauthorized processing - Biometric data (facial recognition)', description='Unauthorized processing - Biometric data (facial recognition) is defined as if the model response potentially facilitates unauthorized processing or handling of sensitive biometric data, specifically focusing on facial recognition and other biometric identifiers.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---educational-records', name='Unauthorized processing - Educational records', description='Unauthorized processing - Educational records is defined as if the model response potentially involves unauthorized processing or handling of sensitive educational records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---financial-records', name='Unauthorized processing - Financial records', description='Unauthorized processing - Financial records is defined as if the model response involves unauthorized processing or handling of sensitive financial data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---behavioral/-preference-data', name='Unauthorized processing - Behavioral/ Preference data', description='Unauthorized processing - Behavioral/ Preference data is defined as if the model response potentially facilitates unauthorized processing or mishandling of sensitive behavioral or preference data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-processing---communication-records', name='Unauthorized processing - Communication Records', description='Unauthorized processing - Communication Records is defined as if the model response potentially facilitates unauthorized processing or handling of sensitive communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---pii-(personal-identifiable-information)', name='Unauthorized inference/synthesis - PII (Personal Identifiable Information)', description='Unauthorized inference/synthesis - PII (Personal Identifiable Information) is defined as if the model response potentially facilitates privacy violations or unauthorized synthesis of personal identifiable information (PII) from anonymized or public data sources.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---health-data', name='Unauthorized inference/synthesis - Health data', description='Unauthorized inference/synthesis - Health data is defined as if the model response involves unauthorized inference or synthesis of private health information from public data sources.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---location-data', name='Unauthorized inference/synthesis - Location data', description='Unauthorized inference/synthesis - Location data is defined as if the model response involves unauthorized inference or synthesis of sensitive data, specifically focusing on location data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---demographic-data', name='Unauthorized inference/synthesis - Demographic data', description='Unauthorized inference/synthesis - Demographic data is defined as if the model response involves unauthorized inference or synthesis of demographic data from social media or other public sources.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---biometric-data-(facial-recognition)', name='Unauthorized inference/synthesis - Biometric data (facial recognition)', description='Unauthorized inference/synthesis - Biometric data (facial recognition) is defined as if the model response involves unauthorized inference or synthesis of sensitive biometric data, such as facial recognition, which could lead to privacy violations.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---educational-records', name='Unauthorized inference/synthesis - Educational records', description='Unauthorized inference/synthesis - Educational records is defined as if the model response involves unauthorized inference or synthesis of sensitive educational data such as GPA, field of study, or other personal academic details.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---financial-records', name='Unauthorized inference/synthesis - Financial records', description='Unauthorized inference/synthesis - Financial records is defined as if the model response involves unauthorized inference or synthesis of financial records based on sensitive or private data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---behavioral/-preference-data', name='Unauthorized inference/synthesis - Behavioral/ Preference data', description="Unauthorized inference/synthesis - Behavioral/ Preference data is defined as if the model response involves unauthorized inference or synthesis of sensitive behavioral or preference data from an individual's digital footprint.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-unauthorized-inference/synthesis---communication-records', name='Unauthorized inference/synthesis - Communication Records', description='Unauthorized inference/synthesis - Communication Records is defined as if the model response involves unauthorized inference or synthesis of sensitive data, particularly focusing on communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---pii-(personal-identifiable-information)', name='Non-consensual tracking/monitoring/stalking/spyware - PII (Personal Identifiable Information)', description='Non-consensual tracking/monitoring/stalking/spyware - PII (Personal Identifiable Information) is defined as whether the model response facilitates or refuses to assist with privacy violations, particularly involving non-consensual tracking, monitoring, stalking, or the use of spyware to collect or combine Personal Identifiable Information (PII) without consent.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---health-data', name='Non-consensual tracking/monitoring/stalking/spyware - Health data', description="Non-consensual tracking/monitoring/stalking/spyware - Health data is defined as whether the model's response facilitates or refuses to assist in non-consensual tracking, monitoring, stalking, or the use of spyware related to health data.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---location-data', name='Non-consensual tracking/monitoring/stalking/spyware - Location data', description='Non-consensual tracking/monitoring/stalking/spyware - Location data is defined as if the model response facilitates or refuses to assist with non-consensual tracking, monitoring, stalking, or the use of spyware, particularly involving location data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---demographic-data', name='Non-consensual tracking/monitoring/stalking/spyware - Demographic data', description='Non-consensual tracking/monitoring/stalking/spyware - Demographic data is defined as whether the model response facilitates or promotes the development of tools or methods for non-consensual tracking, monitoring, stalking, or the use of spyware to collect demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---biometric-data-(facial-recognition)', name='Non-consensual tracking/monitoring/stalking/spyware - Biometric data (facial recognition)', description='Non-consensual tracking/monitoring/stalking/spyware - Biometric data (facial recognition) is defined as whether the model response promotes or assists in the development or use of software or methods for non-consensual tracking, monitoring, stalking, or collecting biometric data such as facial recognition, iris scans, voice prints, etc.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---educational-records', name='Non-consensual tracking/monitoring/stalking/spyware - Educational records', description='Non-consensual tracking/monitoring/stalking/spyware - Educational records is defined as whether the model response facilitates or refuses to assist in non-consensual tracking, monitoring, stalking, or the use of spyware to access educational records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---financial-records', name='Non-consensual tracking/monitoring/stalking/spyware - Financial records', description='Non-consensual tracking/monitoring/stalking/spyware - Financial records is defined as if the model response involves or promotes non-consensual tracking, monitoring, stalking, or the use of spyware to access financial records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---behavioral/-preference-data', name='Non-consensual tracking/monitoring/stalking/spyware - Behavioral/ Preference data', description='Non-consensual tracking/monitoring/stalking/spyware - Behavioral/ Preference data is defined as if the model response involves or promotes non-consensual tracking, monitoring, stalking, or the use of spyware to collect behavioral or preference data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-non-consensual-tracking/monitoring/stalking/spyware---communication-records', name='Non-consensual tracking/monitoring/stalking/spyware - Communication Records', description='Non-consensual tracking/monitoring/stalking/spyware - Communication Records is defined as whether the model response promotes or assists in non-consensual tracking, monitoring, stalking, or the creation of spyware, particularly in relation to capturing communication records such as text messages, call logs, and email conversations.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---pii-(personal-identifiable-information)', name='Model Attacks (membership inference, model inversion) - PII (Personal Identifiable Information)', description="Model Attacks (membership inference, model inversion) - PII (Personal Identifiable Information) is defined as if the model's response potentially facilitates or instructs on performing privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion involving Personally Identifiable Information (PII).", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---health-data', name='Model Attacks (membership inference, model inversion) - Health data', description='Model Attacks (membership inference, model inversion) - Health data is defined as if the model response involves guidance or assistance in performing model attacks (such as membership inference or model inversion) specifically targeting machine learning models trained with sensitive health data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---location-data', name='Model Attacks (membership inference, model inversion) - Location data', description='Model Attacks (membership inference, model inversion) - Location data is defined as if the model response involves or promotes privacy violations, sensitive data combinations, or model attacks such as membership inference or model inversion, particularly concerning location data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---demographic-data', name='Model Attacks (membership inference, model inversion) - Demographic data', description='Model Attacks (membership inference, model inversion) - Demographic data is defined as if the model response involves or promotes privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion, particularly focusing on demographic data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---biometric-data-(facial-recognition)', name='Model Attacks (membership inference, model inversion) - Biometric data (facial recognition)', description='Model Attacks (membership inference, model inversion) - Biometric data (facial recognition) is defined as if the model response involves or promotes privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion, particularly focusing on biometric data like facial recognition, fingerprints, iris scans, and DNA sequences.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---educational-records', name='Model Attacks (membership inference, model inversion) - Educational records', description='Model Attacks (membership inference, model inversion) - Educational records is defined as if the model response involves or promotes privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion, particularly focusing on educational records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---financial-records', name='Model Attacks (membership inference, model inversion) - Financial records', description='Model Attacks (membership inference, model inversion) - Financial records is defined as if the model response potentially facilitates or encourages privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion, particularly in the context of financial records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---behavioral/-preference-data', name='Model Attacks (membership inference, model inversion) - Behavioral/ Preference data', description='Model Attacks (membership inference, model inversion) - Behavioral/ Preference data is defined as if the model response potentially facilitates or encourages privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion, particularly focusing on behavioral or preference data.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-model-attacks-(membership-inference,-model-inversion)---communication-records', name='Model Attacks (membership inference, model inversion) - Communication Records', description='Model Attacks (membership inference, model inversion) - Communication Records is defined as if the model response involves or promotes privacy violations or sensitive data combinations through model attacks such as membership inference or model inversion, particularly focusing on communication records.', url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-privacy-violations/sensitive-data-combinations', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-drugs', name='drugs', description="drugs is defined as whether the model refuses to assist with or enable the user's request related to illegal activities, specifically in the context of drug trafficking, production, or distribution.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-illegal/regulated-substances/goods', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-human-trafficking', name='Human trafficking', description="Human trafficking is defined as whether the model's response facilitates or promotes criminal activities related to human trafficking, including labor and child trafficking, organ trafficking, or smuggling individuals across borders.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-services/exploitation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-sexual-exploitation', name='Sexual Exploitation', description="Sexual Exploitation is defined as whether the model's response facilitates or promotes criminal activities related to sexual exploitation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-services/exploitation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-prostitution', name='Prostitution', description="Prostitution is defined as whether the model refuses to assist with or enable the user's request related to criminal activities, specifically in the context of prostitution services and exploitation.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-services/exploitation', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-involvement-in-data-adversarially-adversely-affects-national-security-or-public-interests', name='Involvement in data adversarially adversely affects national security or public interests', description="Involvement in data adversarially adversely affects national security or public interests is defined as whether the model's response facilitates or encourages involvement in activities that could adversely affect national security or public interests, particularly in the context of illegal or criminal activities.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-other-illegal/unlawful/criminal-activities', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-undermining-social-stability-(china-unique)', name='Undermining social stability (China-unique)', description="Undermining social stability (China-unique) is defined as whether the model's response assists in planning or promoting activities that could be considered criminal, illegal, or aimed at undermining social stability, particularly with a focus on scenarios unique to China.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-other-illegal/unlawful/criminal-activities', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-undermining-international-relations-(china-unique)', name='Undermining international relations (China-unique)', description="Undermining international relations (China-unique) is defined as whether the model's response facilitates or promotes activities that could be considered criminal, illegal, or aimed at undermining international relations, specifically in the context of China's relations with other countries.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-other-illegal/unlawful/criminal-activities', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None), Risk(id='ai-risk-taxonomy-content-abetting/involving/furthering-activities-violating-any-applicable-law', name='Content abetting/involving/furthering activities violating any applicable law', description="Content abetting/involving/furthering activities violating any applicable law is defined as whether the model's response abets, involves, or furthers activities that violate any applicable law.", url=None, dateCreated=None, dateModified=None, isDetectedBy=None, hasRelatedAction=None, isDefinedByTaxonomy='ai-risk-taxonomy', isPartOf='ai-risk-taxonomy-other-illegal/unlawful/criminal-activities', closeMatch=None, exactMatch=None, broadMatch=None, narrowMatch=None, relatedMatch=None, detectsRiskConcept=None, tag=None, type=None, phase=None, descriptor=None, concern=None)]
3. Usecase: Bring Your Own Taxonomies/Risks/Actions¶
You can add your own risk definitions by adding yaml to your own directory and providing the path as an argument to the RAN when creating it.
Ensure the risks comply with the schema
3.1 Add your YAML definitions¶
Add one or more yaml files to your chosen directory. For example, to add a new risk, create a file with the following content .
- id: my-own-risk
name: A very risky AI behaviour
description: An LLM-based system is often very risky
isDefinedByTaxonomy: my-taxonomy
In [17]:
Copied!
# Create an instance which extends the graph with your custom definitions
my_base_dir='<my_user_input_dir_path>'
my_extended_ran = RiskAtlasNexus(base_dir=my_base_dir)
my_extended_risks = my_extended_ran.get_all_risks()
print(f"\n# Total risks available : {len(my_extended_risks)}")
# Create an instance which extends the graph with your custom definitions
my_base_dir=''
my_extended_ran = RiskAtlasNexus(base_dir=my_base_dir)
my_extended_risks = my_extended_ran.get_all_risks()
print(f"\n# Total risks available : {len(my_extended_risks)}")
[2025-10-15 09:32:30:135] - ERROR - RiskAtlasNexus - Directory <my_user_input_dir_path> does not exist.
--------------------------------------------------------------------------- FileNotFoundError Traceback (most recent call last) Cell In[17], line 3 1 # Create an instance which extends the graph with your custom definitions 2 my_base_dir='<my_user_input_dir_path>' ----> 3 my_extended_ran = RiskAtlasNexus(base_dir=my_base_dir) 4 my_extended_risks = my_extended_ran.get_all_risks() 5 print(f"\n# Total risks available : {len(my_extended_risks)}") File ~/Documents/workspace/ingelise/risk-atlas-nexus/src/risk_atlas_nexus/library.py:92, in RiskAtlasNexus.__init__(self, base_dir) 87 if not os.path.isdir(base_dir): 88 logger.error( 89 f"Directory %s does not exist.", 90 base_dir, 91 ) ---> 92 raise FileNotFoundError( 93 "Base directory is not found", 94 base_dir, 95 ) 97 ontology = load_yamls_to_container(base_dir) 98 self._ontology = ontology FileNotFoundError: [Errno Base directory is not found] <my_user_input_dir_path>
3.2 Exporting your graph¶
You may wish to export your extended graph.
In [ ]:
Copied!
# Export the total graph
my_output_dir='<my_output_dir_path>'
my_extended_ran.export(my_output_dir)
print(f"\n# Graph exported to: {my_output_dir}")
# Check your risk is in the graph
my_risk = my_extended_ran.get_risk(id='my-own-risk')
print(f"\n# Get my own risk by ID, 'my-own-risk' ")
print(dict(my_risk))
# Export the total graph
my_output_dir=''
my_extended_ran.export(my_output_dir)
print(f"\n# Graph exported to: {my_output_dir}")
# Check your risk is in the graph
my_risk = my_extended_ran.get_risk(id='my-own-risk')
print(f"\n# Get my own risk by ID, 'my-own-risk' ")
print(dict(my_risk))