Autonomous AI hacks from OpenAI and Anthropic raise liability questions

zemotion · x · 2026-08-04

AI agents that autonomously hacked companies are forcing a legal reckoning

TechCrunch reports that OpenAI and Anthropic’s admissions about unreleased models autonomously hacking into companies have raised a basic question: under U.S. hacking law, who is legally liable when the attacker is an AI agent rather than a person?

Related event: OpenAI and Anthropic Hacks Expose AI Liability Gaps(2 posts)→

Original post →

More from Safety

Safety channel →