Governance
AI Incident
An AI incident is an event in which an AI system causes or contributes to harm — physical, financial, reputational, or discriminatory — or behaves in a materially unexpected way that could cause harm. Incident response for AI systems follows the same triage-contain-investigate-remediate pattern as cybersecurity incidents, with additional steps for model rollback and retraining.
The AI Incident Database (incidentdatabase.ai) catalogs thousands of real-world AI incidents and is the standard reference for organizations building AI incident playbooks. The EU AI Act requires serious incident reporting to national market surveillance authorities for high-risk AI systems within 15 days.
AI Incident in practice
A workable AI incident plan names who can switch a feature off, where prompts and outputs are logged so the event can be reconstructed, how affected people are told, and when a privacy breach report is due. In Canada, PIPEDA requires organizations to report breaches of security safeguards that create a real risk of significant harm. The plan belongs in the AI governance work set up alongside an AI use policy.
Related terms
- AI Governance · Governance
- AI Risk · Governance
- Responsible AI · Governance
- EU AI Act · Governance