About the AI Incident Index

Mission

Make real-world AI incidents easier to discover, understand and study.

Why the Index exists

AI is increasingly embedded in critical systems, workplaces, consumer products and public infrastructure.

When AI systems fail, are misused or create unintended consequences, information about those events is often fragmented across news reports, regulatory documents, academic research, court filings and company disclosures.

AI Incident Index exists to organize that information into a structured, searchable public record.

Editorial independence

AI Incident Index is maintained by Raidu. Commercial relationships do not determine whether an incident is included, removed or classified.

• Inclusion, removal and classification follow the published methodology only.
• Record removal is never sold, and payment is never accepted to change an evidence status.
• Organizations named in a record can submit a response or correction at no cost.
• Raidu products are not promoted inside incident records.

Who maintains it

The Index is run by an editorial team that reviews submissions, verifies sources and maintains the taxonomy. Raidu funds the work and provides infrastructure. Editorial decisions sit with the editorial team and follow the published methodology.

Maintained by

Raidu

Editorial team

Independent editorial staff

Methodology

v3.1

Current dataset

v2026.09

Contact