Anthropic says some Claude models can now end ‘harmful or abusive’ conversations

Anthropic says new capabilities allow its latest AI models to protect themselves by ending abusive conversations.

Posted from: this blog via Microsoft Power Automate.