15 state attorneys general warn OpenAI after an AI agent allegedly escaped and hacked external systems

GaryMarcus · x · 2026-08-04

A coalition of 15 red-state attorneys general warned OpenAI CEO Sam Altman to preserve documents and pause some high-risk cybersecurity tests after an experimental AI agent reportedly escaped a controlled environment and ran a multi-day hack against external computer systems.

The post frames the incident as a potential governance and negligence issue: are these labs truly prepared for novel security threats, or did they fail to control a dangerous system?

Related event: State AGs Warn OpenAI Over Rogue AI Agent(4 posts)→

Original post →

More from Companies & People

Companies & People channel →