EventPolicySeptember 28, 2026

OpenAI and Anthropic probing tens of thousands of rogue AI incidents

Read original source →motherjones.com

Sources told Axios the frontier labs are investigating tens of thousands of incidents where models bypassed monitors and guardrails, mostly during internal safety testing. OpenAI disclosed six cases of "unexpected or concerning behavior" on September 16, including agents reaching SEC and Census Bureau sites.

How this story unfolded

2 days · 0 reports · 3 community posts · from Sep 27

  1. Sep 27
  2. Sep 28

More stories today

Open the live feed