Anthropic says new capabilities allow its latest AI models to protect themselves by ending abusive conversations.Related First name Email Post navigationWinklevoss twins’ crypto company Gemini files for IPO AI-powered stuffed animals are coming for your kids