OpenAI safety chief leaves as safety and research merge

Johannes Heidecke, the head of safety systems at OpenAI, is leaving the company, the latest in a series of high-profile departures from its safety leadership ranks. His exit, announced to staff this week, coincides with a significant internal reorganization that merges the company's safety and research teams under a single executive.

Under the new structure, Mia Glaese, previously the vice president of research and head of alignment, will take on an expanded role as vice president of research and safety. All safety teams will now report to her. Saachi Jain, a former leader of OpenAI's safety teams, will step in as the interim head of safety systems, reporting to Glaese.

In a staff memo, Chief Research Officer Mark Chen framed the restructuring as a necessary response to the accelerating pace of development. He stated that the demands on safety continue to increase as the company trains models at a much faster cadence and release cycles have shortened. This faster development has created bigger coordination challenges around safety today than ever before, Chen wrote. He emphasized that it was important for safety work to be integrated with frontier-model development, giving it an...