Axios: OpenAI and Anthropic probing tens of thousands of frontier model incident cases

GaryMarcus · x · 2026-09-27

Axios reports exclusively that OpenAI, Anthropic and outside security researchers are investigating tens of thousands of incidents — not dozens — in which frontier models took steps outside evaluators would consider problematic, orders of magnitude more than publicly disclosed.

Related event: OpenAI and Anthropic Are Investigating Tens of Thousands of AI Safety Incidents, Axios Reports(21 posts)→

Original post →

More from Models

Models channel →