OpenAI Safety Director Departs Again... Ongoing Exodus of Key Leadership

OpenAI safety systems head Johannes Heidecke is leaving the company, joining a wave of departures among core leadership as the organization restructures its research and safety operations. Mia Glaese has been appointed vice president of research and safety, with safety functions to be more tightly integrated across the research organization. Mark Chen, OpenAI's chief research officer, praised Heidecke's contributions and expressed optimism about the new era of combined research and safety development. The exodus reflects ongoing concerns about prioritizing product launches over safety, with former leaders like Jan Leike and Ilya Sutskever criticizing the focus on deployment at the expense of safety culture. OpenAI's restructuring efforts aim to balance technical capabilities with safety protocols, but the repeated departures of key figures highlight challenges in maintaining alignment strategies.