Anthropic rolled out a feature letting its AI assistant terminate chats with abusive users, citing "AI welfare" concerns and sparking debates about digital consciousness. https://decrypt.co/335732/claude-rage-quit-conversation-own-mental-health