Glossary
AI Incident Reporting
The practice of documenting and sharing information about AI system failures, misuse or harms so that others can learn and respond.
Last reviewed 2026-08-22
In plain language
Incident reporting exists in aviation, medicine and cybersecurity. For AI, official reporting duties appear in some regulations for serious incidents, while independent databases collect public cases. Definitions of “incident” vary. Reporting only works if organizations are required or incentivized to disclose, and if reports reach people who can act.
Why it matters for AI governance
Without shared incident knowledge, risk management is anecdotal. Reporting is a public-good infrastructure for AI governance.