Risk events that matter to specialty insurance
AI-powered event intelligence with automated detection, classification, and transparent review status

Correction Policy

RiskEvents is committed to accuracy. Our AI classification pipeline may produce errors in event categorisation, impact assessment, geographic attribution, or entity identification. We actively monitor classification quality and welcome correction requests from users.

Reporting errors

If you identify an error in any event listing, please submit a correction request via our contact form. Include the event title or URL, the specific error, and if possible, a source supporting the correct information.

How corrections are processed

Correction requests are reviewed by our team. When a correction is applied, the event's review status changes to “Human Verified” and a structured correction record is created, capturing what changed, the before and after values, and the source of the correction.

Correction data is used to measure and improve AI classification quality over time. We track correction rates by field, peril category, and source type to identify systematic issues.

Retractions

In cases where an event is determined to be fundamentally incorrect (e.g., based on a hoax, misidentification, or retracted source reporting), the event may be retracted entirely. Retracted events remain visible in the Historical view with a “Retracted” status label. Subscribers who received alerts for the event will receive a retraction notification.

Event removal

In exceptional circumstances, events may be removed from public view for legal, privacy, or safety reasons. Removed events appear as placeholders in the Historical view without identifying details.

Response time

We aim to acknowledge correction requests within one working day and to resolve them within three working days. Urgent corrections (e.g., events that may cause harm if left uncorrected) are prioritised.