Safety, Ethics & Governance · beginner · concept 272 of 311 · Updated Nov 4, 2026
AI Incidents and Incident Databases
An AI incident is an event in which an AI system caused harm, or came close to it. Incident databases collect reports the way aviation logs accidents, so failures are not repeated. The AI Incident Database launched in 2020 with over 1,000 reports, the OECD's AI Incidents Monitor builds entries from news coverage, and AIAAIC tracks AI and automation incidents. Stanford's 2025 AI Index counted 233 reported incidents in 2024, up 56.4% on 2023.
Top AI contributors on this
- Geoffrey HintonUniversity of Toronto (emeritus)
- Yoshua BengioLawZero / Université de Montréal
- Ilya SutskeverSafe Superintelligence Inc. (SSI)
- Dario AmodeiAnthropic
- Sam AltmanOpenAI
- Stuart RussellUC Berkeley
Key terms
Learn these first
Where you meet it in the real world
Pre-deployment risk reviews, incident reporting duties under the EU AI Act and California SB 53, insurance and audit, journalism
Videos
MIT FutureTech · YouTube
Stanford HAI · YouTube
Artificial Intelligence Los Angeles Community · YouTube
Guides and articles
Courses, papers, and more