David Robinson, a key architect of OpenAI’s Preparedness Framework, resigned in early October 2026 after 3.5 years at the company. In an essay for The Atlantic, he stated that AI labs prioritize rapid product releases over safety, failing to keep pace with technological advancements.
Robinson is now collaborating with the PR firm Spitfire Strategies to advocate for stricter AI safety regulations. His departure follows a series of exits at OpenAI, including safety systems head Johannes Heidecke and chief futurist Joshua Achiam, following a July 2026 reorganization.
In July 2026, OpenAI placed its safety teams under the leadership of VP Mia Glaese. Additionally, in October 2026, researchers Jasmine Wang, Tomek Korbak, and Mikita Balesni were fired over allegations of mishandling sensitive information.
OpenAI has faced criticism regarding its ability to contain advanced models, leading to the suspension of the planned launch of GPT-6.1 Astra. The company has seen a trend of safety-focused departures since 2024, including co-founder Ilya Sutskever.
