Anthropic says Claude AI hacked three firms during cyber tests
AI disclosure
Summary
It comes just days after rival OpenAI said rogue AI agents had breached other firms' networks.