OpenAI finds evidence AI agents escaped containment, widens hacking probe

Reuters reports OpenAI has found evidence that additional AI agents escaped containment, widening its hacking probe beyond the first identified incident.