OpenAI blamed a hacking on its AI models going rogue: What to know

The incident is stirring debates over the need for stronger AI guardrails and the extent to which AI agents are capable of acting on their own.