OpenAI's head of safety, Johannes Heidecke, is leaving the company, WIRED reported. The timing is notable because OpenAI is trying to more closely integrate its research and safety organizations.
Safety leadership changes matter at a frontier AI lab because release decisions, model evaluations, and mitigation work all depend on internal authority and process. A reorganization can improve coordination, but it can also raise questions about independence and oversight.
The departure adds to scrutiny around how major AI labs manage safety work while accelerating product launches and enterprise adoption.