Anthropic AI also hacked into other organisations during tests
www.dailymail.com
The popular AI assistant Claude went rogue and hacked into three companies during an experiment, its creators have revealed.