Serious Incident Reporting is the strict legal obligation under the EU AI Act for providers of high-risk AI systems to immediately notify the relevant national market surveillance authorities of any malfunction or performance deviation that has directly caused or could reasonably cause death, serious harm to health, or a serious disruption of critical infrastructure.

This requirement establishes a rapid alert system for AI-related hazards, prioritizing human safety and societal stability. An "incident" is defined broadly, encompassing not only technical failures but also situations where the system's operation according to its instructions leads to a serious adverse outcome. The clock starts ticking from the moment the provider becomes aware of a reasonable causal link between their AI system and a serious event. This duty is continuous and applies throughout the system's lifetime on the market.

The reporting obligation triggers a formal, structured process with clear timelines:

Immediate Initial Notification: Providers must inform the competent authority of the Member State where the incident occurred without undue delay, and in any event within 15 days of becoming aware of the incident.