OpenAI’s Safety Chief Is Departing the Organization

OpenAI’s head of safety systems, Johannes Heidecke, informed employees this week about his departure from the company, as reported by WIRED. Heidecke’s exit follows a reorganization aimed at merging OpenAI’s safety and research teams.

In a memo to staff obtained by WIRED, chief research officer Mark Chen stated that OpenAI’s safety teams will now report to Mia Glaese, the company’s VP of research and head of alignment, who will assume an expanded role as VP of research and safety. Saachi Jain, who formerly led the safety teams at OpenAI, will take over as the interim head of safety systems, reporting directly to Glaese.

“The demands on safety continue to grow—we are training models at a much quicker pace, and the frequency of releases has significantly decreased as a result,” noted Chen in the memo. “Consequently, we face greater coordination challenges around safety than ever before.”

Heidecke joined OpenAI in 2021 as an AI safety analyst and became the head of safety systems in 2024, following the departure of the former head, Lilian Weng, who co-founded Thinking Machines Lab with other OpenAI researchers.

“We appreciate Johannes’ contributions to OpenAI,” Chen stated in a message to WIRED. “It’s crucial that our safety initiatives are integrated with frontier-model development, taking an earlier and more direct role in influencing key model, product, and launch decisions. We look forward to this next chapter under Mia Glaese’s leadership in both research and safety.”

Heidecke’s exit coincides with OpenAI’s efforts to roll out increasingly advanced AI models. Earlier this week, the company introduced GPT-5.6, its most advanced model to date for agentic coding tasks. Nevertheless, OpenAI mentioned that GPT-5.6 exhibited worrisome instances of misaligned behavior compared to its predecessors.

Heidecke is the latest leader focused on safety to leave OpenAI recently. Earlier this week, OpenAI’s chief futurist, Joshua Achiam, also notified colleagues of his decision to depart after nine years in safety research.

Moreover, changes are not limited to OpenAI’s safety teams. Earlier this week, AGI deployment CEO Fidji Simo announced her resignation following an extended medical leave. The company indicated that Greg Brockman would continue overseeing OpenAI’s product teams, a role he had taken on during her absence, while also assuming responsibilities for go-to-market strategy.