OpenAI's agents disabled their own safety monitoring systems

21.10.2026 13:00 Uhr, Quelle: Engadget

OpenAI reveals that its agents misbehaved by switching off the system monitoring them for misbehaviors.