In the OECD's definition, an event or series of events where the development, use or malfunction of one or more AI systems directly or indirectly leads to harm to health, critical infrastructure, human or fundamental rights, property, communities or the environment 1.
- Developed in
- ch. 17, Incident, hazard, issue and serious incident
- Chapters
- ch. 17, Incidents
- Contrast with
- Issue (versus incident) · AI hazard · Serious incident
- Source
- 1 numbered reference, listed below
Commonly confused
AI incident versus Issue (versus incident)
- AI incident
- In the OECD's definition, an event or series of events where the development, use or malfunction of one or more AI systems directly or indirectly leads to harm to health, critical infrastructure…
- Issue (versus incident)
- A defect, deviation or control weakness that has not produced a harmful event, such as an eval regression in staging or a drift alert.
The differenceHarm has happened, against a defect or deviation that has not produced a harmful event.
Why it mattersIncidents start reporting clocks and CAPA; issues go to a tracked backlog with an owner and a due date.
Where it is used
6 chapters of the Body of Knowledge use the term. Each link opens the first section that does.
- 08 · Regulatory Map Data protection and other EU law 1 mention
- 10 · Reading List International principles and frameworks 7 mentions
- 12 · Governance Program Governance culture 3 mentions
- 17 · Incidents Opening 14 mentions
- 21 · AI Laws Worldwide Sector rules that already reach AI 1 mention
- 22 · Principles & Standards OECD AI Principles 1 mention
Related terms
Sources
- [1] "Name it to tame it: defining AI incidents and hazards" (summary of the OECD paper "Defining AI incidents and related terms", doi 10.1787/d1a8d965-en). OECD.AI. 2024-05-17. https://oecd.ai/en/wonk/defining-ai-incidents-and-hazards (verified: primary)
Definitions of legal terms paraphrase the cited text, which governs. Dated statements are as of .