risk
AI safety failures causing physical or psychological harm
AI errors in safety-critical systems (autonomous vehicles, medical devices, industrial controls) cause injury or death; safety-constraint violations by agentic AI, cascading failures across coupled systems, and AI-generated misinformation/deepfakes cause harm.
In catalog since 2026-09-17T22:28:00Z · Last changed 2026-09-17T22:28:00Z (f368a6cce277)
Record JSON · Open in map · Data retrieval guide
Catalog revision: 791ff2dd3a45707290badee660f185e514d15f1cf518908628f425c2f2c56ee4. A connection does not establish full coverage.
Attributes
- category
- ai_governance
- domain
- AI Governance
- Risk Assessment & Management
- taxonomy
- nist-ai-rmf-risk
- iso-23894-ai-risk
- inherent_rating
- high
Details
- risk_id
- ai-safety-harm-to-people
- category
- ai_governance
- likelihood
- low
- impact
- critical
- inherent_rating
- high
- treatment
- mitigate
- taxonomies
- nist-ai-rmf-risk
- iso-23894-ai-risk
Source
No record-specific source URL is provided.
Connections
- UC-RISK-01 — Establish and maintain a tailored risk management framework mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- AI-Act Art 9 lifecycle risk processes in the framework establish the governance under which high-risk-AI safety risks are identified and mitigated.
- UC-AI-19 — Constrain agent actions and tool use to authorized scope mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Human approval before high-impact actions keeps an agent from causing physical or financial harm autonomously.
- UC-AI-07 — Verify, validate, and control AI deployment and changes mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Verifying safety requirements before release blocks unsafe systems from shipping.
- UC-AI-15 — Fulfill general-purpose AI model provider obligations mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Assessing/mitigating systemic risks and reporting serious incidents reduces large-scale harm from powerful models.
- UC-AI-23 — Prevent misuse of AI systems for cyber offense and catastrophic harm mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Refusing uplift for weapons and attacks prevents the most severe category of AI-enabled harm to people.
- UC-AI-11 — Operate AI concern, incident, and external reporting channels mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Incident channels detect and communicate safety harms to affected parties for response, reducing impact.
- UC-AI-08 — Log and monitor AI system behavior in operation mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Behavior monitoring with alerting detects unsafe operation for escalation, reducing harm impact.
- UC-AI-16 — Ensure human oversight of AI decisions mitigates AI safety failures causing physical or psychological harm
- strength
- primary
- rationale
- Human ability to override or halt a malfunctioning system is the last-line operational defense against safety-critical AI harm.
- UC-AI-20 — Prevent harmful, out-of-scope, hallucinated, and over-exposed AI outputs mitigates AI safety failures causing physical or psychological harm
- strength
- primary
- rationale
- Output filtering against the harmful-output taxonomy is the last technical barrier before a harmful response reaches a person.
- UC-AI-12 — Enforce responsible and lawful use of AI systems mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Restricting systems to their intended use reduces misuse-driven harm.
- UC-AI-05 — Set responsible AI development objectives and requirements mitigates AI safety failures causing physical or psychological harm
- strength
- related
- rationale
- Embedding safety objectives in the design/development process reduces unsafe design choices before build begins.