OECD AI Incident Reporting Framework
Common 29-criteria framework designed to support interoperable reporting and monitoring of AI incidents across jurisdictions and sectors.
Common 29-criteria framework designed to support interoperable reporting and monitoring of AI incidents across jurisdictions and sectors.
Governments, authorities and other stakeholders establishing or aligning incident-reporting processes.
Framework can support voluntary or mandatory reporting schemes but is not itself a statutory reporting requirement.
Common 29-criteria framework designed to support interoperable reporting and monitoring of AI incidents across jurisdictions and sectors.
Track hazards that could plausibly lead to material AI incidents, not only realised harm.
Actor: AI actors/authorities · Domain: monitoring
Evidence: Hazard register, trend analysis and escalation rules.
Capture consistent information about system, context, harm, affected actors and incident circumstances.
Actor: AI actors/authorities · Domain: incident response
Evidence: Incident record with dates, system context, harm taxonomy, source and response.
Cybatar's mapping is designed for AI governance and assurance work. Legal interpretation remains anchored to the current primary text and competent authority guidance.
Open OECD source