Annual AI Threat Reports
Analytical reports on documented AI-enabled threat incidents — with domain breakdowns, severity trends, and interpretive analysis.
For researchers, policymakers, and security teams tracking the evolving landscape of real-world AI failures. All figures computed from the incident database at build time.
Year-to-Date Report
85 incidents across 8 domains so far. 22 critical.
Domain analysis, severity trends, failure stage breakdown, top threat patterns, and all documented incidents with interpretive analysis.
Annual Report
40 incidents across 7 domains. 12 critical.
Full-year analysis with domain distribution, severity breakdown, failure stages, top threat patterns, and all documented incidents.
Annual Report
25 incidents across 7 domains. 6 critical.
Full-year analysis with domain distribution, severity breakdown, failure stages, top threat patterns, and all documented incidents.
How reports work: All quantitative data is computed from the incident database at build time. Interpretive analysis is grounded in the data and follows the 8-domain taxonomy.
Reports are updated with each site build as new incidents are added to the database.