What this site is — and how it's built.
AIIncidentTracker is an editorial database of documented AI incidents — hallucinations, accidents, discrimination, fraud, and safety failures involving deployed AI systems. Every entry is grounded in primary sources, every claim has a hedge-language wrapper until it is judicially or regulator-confirmed, and every record carries a last-verified date. Updated daily.
What we are
AIIncidentTracker covers documented incidents involving AI systems — LLM hallucinations, autonomous-vehicle accidents, algorithmic hiring bias, facial-recognition misidentifications, deepfake fraud, prompt-injection attacks, recommender-system harms, and more.
Our audience is compliance officers, plaintiff-side attorneys, legal-tech journalists, GRC teams, AI safety researchers, and insurance underwriters writing AI liability policies.
Everything we publish is grounded in primary sources — court filings, regulator reports, company statements, peer-reviewed research, and verifiable mainstream news. Standards are documented in full on the Methodology page: primary sources only, hedge language mandatory, every claim sourced, and a multi-stage review before any record is published.
Editorial team
AIIncidentTracker is produced by an editorial team focused exclusively on documented AI harms. The team's qualifications, conflict-of-interest policy, and review pipeline are detailed on the Editorial standards page.
Editor bios with credentials and links will appear here before public launch. Soft-launch builds publish only to a beta cohort while named editors are onboarded and the legal-review process is established.
How we're funded
- Pro subscriptions. Pricing has not been announced yet. When it is, the structure and price points will be published on /pro. Pro will add daily alerts, advanced filters, full data export, and API access on top of the free database.
- Newsletter sponsorships. Clearly labeled slots in the weekly Incident Brief, capped per issue, and excluded from incident coverage.
- Vendor referrals. When an incident matches a relevant vendor category, we cross-link to AIComplianceVendors. Some of those vendors pay a referral fee — appearance on an incident page is determined by failure-mode matching, never by commercial relationship.
- Data licensing. Selective access provided to academic researchers and accredited media organizations.
What we don't do
- We do not give legal advice. Anything on this site is informational only.
- We do not source from press releases, social media without follow-up reporting, or unverified summaries.
- We do not run paid placement on incident prominence or company profiles.
- We do not editorialize on active incidents — we state facts, link to sources, and let readers draw conclusions.
- We do not publish incidents that require identifying minors to tell.
- We do not embed sponsored content inside incident coverage.
Contact
- Editorial / corrections: editorial@aiincidenttracker.com
- Removal / correction requests: /removal-request
- Press inquiries: /press
- Methodology questions: /methodology
- General: /contact
Founding editorial team
Bylines and headshots appear here once editors are onboarded through Clerk and the /admin/editors panel. Until then, all editorial decisions on this site are attributable to the founding editorial team, who can be reached at editorial@aiincidenttracker.com.