OpenAI’s head of security systems, Johannes Heidecke, is leaving the company following an internal restructuring that brought together the security and research teams under a single leader, Wired reported Friday. Chief Research Officer Mark Chen said in a memo to staff that the security teams will now report to Mia Glaese, whose title has been expanded to VP of Research and Security, a newly created role.
Saachi Jain has been named interim head of security systems while the company searches for Heidecke’s permanent replacement. This is the second time in less than two years that OpenAI has incorporated a security organization into a structure that reports to a research chief.
Heidecke’s tenure
Heidecke joined OpenAI in 2021 as an AI security analyst and succeeded Lilian Weng as head of security systems in 2024. His work included model alignment, rules-based reward systems, and company readiness assessments for potentially hazardous model opportunities.
💜 of EU technology
The latest rumblings from the EU tech scene, a story from our wise founder Boris and some questionable AI art. Free in your inbox every week. Register now!
Chen thanked Heidecke in a memo to staff, saying it was “important to integrate our security business with frontier model development, which has an earlier and more direct role in shaping core model, product and sales decisions.” Heidecke is the latest in a string of top security executives to leave or reshuffle the company over the past two years.
Dispersion pattern
The Superalignment team, which OpenAI announced pledged 20% of the company’s computing in 2023, was disbanded in May 2024 after co-chairs Ilya Sutskever and Jan Leyke left. When Leike left, he wrote publicly that “safety culture and processes took a backseat to brilliant products.”
The AGI Readiness team followed in October 2024 when its leader, Miles Brundage, resigned; The Mission Alignment team, the successor to Superalignment, disbanded in February 2026 after 16 months, and its leader Joshua Achiam took the new title of “chief futurist”. In April, OpenAI lost its head of product, head of Sora, and CTO of the enterprise in one day.
Heidecke’s departure comes as more senior leaders continue to depart. Fidji Simo, Head of Applications at OpenAIresigned this month due to a lengthy medical recovery.
Work for integration
Glaese’s expanded title suggests that OpenAI wants security to remain a named priority even in the new structure. The company launched the Safety Scholarship on April 6inviting outside researchers to conduct independent safety and compliance work in the laboratory.
The rationale, expressed by Chen, is that placing safety within research gives it a place in model decisions from the start, rather than as a last checkpoint before starting work. Critics argue that a safety group reporting within a study has less structural independence than a separate reporting one and less leverage to delay or block a product.
External pressure
The move comes as OpenAI goes with enhanced external validation. Attorneys general of 42 states have started investigations filed a subpoena shortly after filing for a stock exchange listing, including the company’s confidential advertising, user data and internal policies.
Lilian Weng, who held a security systems role before Heidecke, has joined Thinking Machines Labs, an AI startup founded by former OpenAI CTO Mira Murati. Murati then warned the public artificial intelligence management lags behind model capabilities.
