Serious Incident (AI Definition)

Under the EU AI Act, a serious incident is an incident or malfunctioning of an AI system that directly or indirectly leads to:

1. Death or serious harm to a person’s health
2. Serious and irreversible disruption of critical infrastructure management
3. Infringement of EU law obligations protecting fundamental rights
4. Serious harm to property or the environment

Both providers and deployers must report serious incidents to authorities without undue delay.

The definition encompasses both direct causation (the AI made a harmful decision) and indirect causation (the AI’s failure led to harmful conditions). It also includes malfunctioning, not just intended behavior gone wrong, but systems not working as designed.

The inclusion of fundamental rights infringement is notable. Harm isn’t limited to physical safety, systematic rights violations through AI systems count as serious incidents requiring reporting.

Related:, 07-molecule—value-chain-accountability-ai, 05-molecule—risk-based-ai-classification