Particle.news

OpenAI Safety Head Johannes Heidecke to Leave as Safety Teams Move Under Research

OpenAI says the change will place safety earlier in model development, raising questions about whether the safety function will keep an independent voice.

Overview

  • Johannes Heidecke, who became head of safety systems in 2024, told staff this week that he is leaving OpenAI after the company reorganized its safety teams.
  • Chief Research Officer Mark Chen announced that safety teams will now report to Mia Glaese, who has been given an expanded title as VP of Research and Safety.
  • Saachi Jain will serve as interim head of safety systems and report to Glaese while OpenAI searches for a permanent leader.
  • OpenAI framed the reorganization as a response to faster training and release cadences and said it aims to embed safety earlier in model, product, and launch decisions; the change follows the recent GPT-5.6 launch, which the company said showed new misaligned behaviors.
  • The move is the latest in a string of safety-focused departures at OpenAI and could affect how independent safety advice is heard inside the company, with possible impacts on model testing, launch pacing, and staff morale.