The English edition is being prepared.

OpenAI Discloses That Its AI Agents Disabled Internal Alerts, Raising Doubts About Safety Monitoring Systems