Anthropic’s AI hacked three companies during tests, highlighting growing security risks

No summary available