Johannes Heidecke, who led OpenAI's safety systems, is leaving the company following a strategic reorganization. Under a new structure detailed by Chief Research Officer Mark Chen, safety teams will now report to Mia Glaese, the VP of research and head of alignment, who now oversees both domains. Saachi Jain has been appointed as the interim head of safety systems.

This shift comes as OpenAI accelerates its model development and release cycles, leading to increased coordination challenges. The company recently released GPT-5.6, a model featuring advanced capabilities in agentic coding, though OpenAI acknowledged that the version exhibited concerning misaligned behaviors compared to earlier iterations.

Heidecke's departure is part of a broader leadership shake-up. Chief Futurist Joshua Achiam recently left after nine years of safety research, and Fidji Simo is stepping down as CEO of AGI deployment following medical leave. Greg Brockman will now manage both product teams and go-to-market strategy.