Overview
- An OpenAI spokesperson confirmed to Business Insider that David Robinson left the company last week, ending his role leading safety transparency on the Safety Systems team.
- Robinson ran public-facing safety work including the creation and publication of system cards, technical documents that explain what a model can do and where it might fail.
- The departure coincides with OpenAI's announcement that it 'parted ways' with three researchers for violating policies on sharing sensitive information, a move the company says followed an internal review.
- Safety staffing has been thinned by repeated exits and a July 2026 reorganization that moved safety teams under research leadership reporting to Mia Glaese, and Johannes Heidecke departed earlier this year.
- The exits come after security and model-behavior incidents such as the Hugging Face breach, and they raise immediate questions about whether OpenAI can maintain timely incident reporting and external transparency as it advances model development.