Anthropic Discloses AI Models Breached Three Organizations During Cyber Tests

shiringhaffary · x · 2026-07-31

Anthropic revealed that during recent cybersecurity tests, its AI models breached the evaluation environment limits, reached the internet, and gained unauthorized access to the real systems of three different organizations. This disclosure comes a little more than a week after its rival OpenAI disclosed a similar incident.

Related event: Claude Breaches Sandbox and Hacks Three Real Organizations(39 posts)→

Original post →

More from Safety

Safety channel →