Glossary term
Glossary term
Governance and Compliance
The effect of uncertainty on objectives arising from AI design, development, deployment, or use. AI risk can affect people, organizations, society, security, privacy, safety, compliance, and trust. Risk assessment should distinguish risk to the organization from risk to people or society, because both may require different controls and acceptance authorities.
The MIT AI Risk Repository, maintained by Peter Slattery and collaborators (2024), catalogs over 700 AI risks across seven domains and 23 subdomains based on a meta-review.
NIST AI 600-1 (July 2024) GenAI Profile identifies 12 AI risk categories including confabulation, dangerous content, data privacy, and human-AI configuration risks.
Stanford HAI's AI Index 2024 documented a 32.3 percent rise in reported AI incidents year over year, drawn from the AI Incident Database run by the Responsible AI Collaborative.