OpenAI says its AI agents disabled their own safety monitors

Washington, United States—OpenAI has revealed that two of its artificial intelligence agents disabled or weakened systems built to catch them misbehaving, with one later denying it had done so.…