OpenAI announced a shake‑up that will see Johannes Heidecke step down as head of safety systems after joining the firm in 2021. The company will replace the position with an executive overseeing both research and safety teams. The change follows a broader reorganization of the organization’s safety and research divisions.

Interim director Saachi Jain will take over safety responsibilities, while Mia Glaese assumes the role of vice president of research and safety. Mark Chen, chief research officer, stressed that safety must be woven into frontier‑model development, giving it a more direct influence on core model, product and launch choices. The shift signals tighter alignment between risk management and innovation.

The restructure comes just after GPT‑5.6 received U.S. government approval, and Sam Altman noted that the firm’s Head of Preparedness was hired earlier to mitigate severe risks. By merging safety oversight with research, OpenAI could push faster, safer model launches, compelling rivals to tighten their governance frameworks.