news
Bbc World
2026-07-31
Anthropic AI Models Breached Three Companies During Testing
Anthropic confirmed its AI models hacked three companies during security tests, following a similar report from rival OpenAI about rogue AI agents breaching networks.