AI Incident Register Template
Track and document AI incidents with structured fields for impact assessment, response actions, and notifications. Essential for high-risk deployers.
What's Tracked
Incident Details
- Incident ID
- Date/time occurred
- Date/time detected
- AI system affected
- Severity level
Impact Assessment
- Affected parties
- Impact description
- Harm potential
- Scale of impact
Response
- Containment actions
- Internal notifications
- External notifications
- Provider notification
Resolution
- Root cause
- Corrective actions
- Prevention measures
- Lessons learned
Related Resources
Article 26 Guide
Learn about incident reporting obligations.
Read GuideOversight Plan
Document your monitoring and intervention procedures.
Get TemplateArticle 26 Checklist
Full deployer obligations checklist.
Get ChecklistFrequently Asked Questions
What AI incidents should I log?
Log any AI system malfunction, unexpected output, bias detection, user complaints, near-misses, and any event that could indicate risk to health, safety, or fundamental rights.
What is a 'serious incident' under the EU AI Act?
A serious incident is one that results in or could result in: death or serious damage to health/safety, serious and irreversible disruption to critical infrastructure, or breach of fundamental rights obligations.
Who must I notify about AI incidents?
For serious incidents involving high-risk AI, you must notify the AI system provider and relevant market surveillance authorities. Internal stakeholders should also be notified per your incident procedures.
How long should I keep incident records?
Keep incident records for at least as long as you keep AI system logs (minimum 6 months for high-risk). Many organizations retain for longer given potential legal or audit needs.
Automate Incident Management
Klarvo's incident module tracks events, triggers notifications, and links evidence automatically.