Head of OpenAI WIRED has learned that Johannes Heidecke, head of safety systems, told staff this week that he was leaving the company. Mr. Heidecke’s departure follows an organizational reorganization aimed at integrating OpenAI’s safety and research teams.
In a memo to staff seen by WIRED, chief research officer Mark Chen said OpenAI’s safety team will now report directly to Mia Grace, the company’s vice president of research and head of alignment, who will take on an expanded role as vice president of research and safety. Saachi Jain, who previously led the safety team at OpenAI, will become interim head of the company’s safety systems division, reporting to Glaese.
“Safety demands continue to rise. We are training models at a much faster pace, and our release cycles are correspondingly shorter,” Chen said in the memo. “As a result, today we face greater coordination challenges than ever before when it comes to safety.”
Mr. Heidecke joined OpenAI in 2021 as an AI Safety Analyst. He became the company’s head of safety systems in 2024, after former head Lilian Wen left to co-found Thinking Machines Lab with other OpenAI researchers.
“We are grateful for Johannes’ contributions to OpenAI,” Chen said in a statement to WIRED. “It is important that our safety efforts are integrated with Frontier model development and that critical model, product and launch decisions are made early and directly. We look forward to taking on this next chapter under Mia Grace’s leadership across research and safety.”
Heidecke’s resignation comes as OpenAI seeks to launch increasingly sophisticated AI models. Earlier this week, the company announced GPT-5.6, its most capable model to date for agent coding tasks. However, according to OpenAI, GPT-5.6 exhibited a concerning form of inconsistent behavior compared to previous models.
Heidecke is a safety-focused leader who left OpenAI in recent days. Earlier this week, OpenAI’s chief futurist, Joshua Achiam, also told colleagues he was leaving the company after nine years in safety research.
And OpenAI’s safety team isn’t the only thing changing. Earlier this week, OpenAI’s CEO of AGI implementation, Fiji Simo, informed staff that he would be stepping down from his position after taking an extended medical leave. The company announced that Greg Brockman will continue to lead OpenAI’s product team and will also be responsible for go-to-market strategy, which was assigned in her absence.
