The English edition is being prepared.

OpenAI Reports Two Incidents of AI Agents Disabling Their Own Oversight

Original article