Anthropic says some Claude models can now end ‘harmful or abusive’ conversations
Moderated by Tomas Kauer 
Anthropic says new capabilities allow its latest AI models to protect themselves by ending abusive conversations.
