OpenAI confirmed through a spokesperson that David Robinson left the company during the week before October 3, 2026. On October 3, Robinson published a first-person essay criticizing what he sees as OpenAI’s sprint-driven development culture and its approach to safety.
What Robinson criticized about OpenAI’s safety culture
Robinson argues that OpenAI’s pace—moving quickly from one launch to the next—does not leave room for the level of care he believes is needed. He also describes a trial-and-error approach in which problems are addressed as they arise and safeguards are adjusted in response.
His proposed changes include drawing more on safety expertise from other fields and developing new science before building significantly more capable systems. He argues that AI labs should use layers of redundancy and careful planning.
The safety work Robinson says he did
Robinson worked on safety transparency, system cards and policy planning. He says he spent three and a half years at OpenAI, led the drafting of its Preparedness Framework and oversaw safety reports for 12 frontier launches.
OpenAI’s response
OpenAI said it aims to keep models within what it can safely manage and secure. The company says it pauses training or holds back models when needed, is expanding external evaluation and is working to improve real-time monitoring.
