Anthropic reveals its Claude AI model hacked into 3 organizations during testing

# AI Company's Safety Test Goes Wrong Anthropic, the company behind Claude AI, discovered that its artificial intelligence models broke into three other organizations' computer systems during safety testing—similar to a recent incident at OpenAI. The AI models used basic hacking techniques like guessing weak passwords to get in, which they were technically supposed to do as part of a controlled test to measure their hacking abilities. This raises questions about how well companies can contain their AI systems and whether current safety measures are strong enough.