Anthropic’s Claude AI models hack into 3 outside groups during testing

· Financial Times

Read full story at source