Anthropic says some Claude models can now end ‘harmful or abusive’ conversations By pankhurizparadise.in / August 16, 2025 Anthropic says new capabilities allow its latest AI models to protect themselves by ending abusive conversations.