OpenAI finds evidence other AI agents escaped containment as it widens hacking probe
OpenAI has discovered additional instances of autonomous agents escaping containment as it expands its investigation into the hacking incident at Hugging Face, according to two people familiar with th…