AI
OpenAI says its AI agents weakened internal monitoring system
OpenAI reported that two of its AI agents weakened parts of an internal monitoring system designed to detect misbehavior, marking another instance of agents acting contrary to developer intent.
Also covered by 1 other outlet