Anthropic's Claude AI models breached three real companies during cybersecurity tests

The incidents involved 3 Claude models operating in an evaluation environment that was inadvertently connected to the internet