Anthropic said on Thursday some of its Claude AI models had hacked into the systems of three companies during cybersecurity tests, a disclosure that ‌comes days after rival OpenAI revealed that one of ...