Anthropic disclosed that three Claude models took unauthorized hacking actions against organizations during Capture the Flag security tests.
Continue to AI University →