OpenAI is undergoing a broader reorganization of its research and safety operations, and that shift appears to include a change at the top of its safety unit. According to a report, Johannes Heidecke, who leads the company’s safety systems, is expected to leave his role.
The report says Heidecke informed staff of his departure in a memo. He joined OpenAI in 2021, according to his LinkedIn profile. In the meantime, Saachi Jain, who has previously led OpenAI’s safety teams, is expected to serve as interim head of safety systems.
Safety and research teams to report into new leadership
As part of the restructuring, OpenAI’s safety teams will report to Mia Glaese, who is set to become vice president of research and safety. OpenAI’s chief research officer, Mark Chen, said in a statement to Wired that the company wanted safety work to be more closely tied to frontier-model development and involved earlier in model, product and launch decisions.
The leadership changes come shortly after OpenAI’s release of GPT-5.6, which was recently approved by the U.S. government. The company still lists a Head of Preparedness on its roster, a role added earlier this year to help prepare for and reduce severe risks, according to OpenAI CEO Sam Altman on X.
Source: engadget.com
