About
David G. Robinson is a writer and technology-policy researcher who spent three and a half years at OpenAI before resigning in late September 2026. A Princeton philosophy graduate and Rhodes Scholar who later took a law degree at Yale, he helped found Princeton's Center for Information Technology Policy and co-founded Upturn, a nonprofit that works with civil rights groups wherever software comes to mediate housing, hiring or criminal justice. He helped start the ACM FAccT conference, worked on the Blueprint for an AI Bill of Rights at the White House Office of Science and Technology Policy, and wrote 'Voices in the Code' (2022), on how the U.S. kidney allocation algorithm was governed in public. He joined OpenAI in May 2023 to build its policy planning team, then moved into the safety team as what he calls its technical translator, leading the writing of the system cards that explain why each release is considered safe. He arrived skeptical of catastrophic-risk arguments; he left, he wrote in The Atlantic, convinced the industry's culture is not safe enough for what it is now building.
Key Contributions
- Co-founded and led Upturn (2011–2020), a nonprofit that partners with civil rights groups on technology in housing, hiring and criminal justice
- Helped found the ACM FAccT conference, and was program chair the year it published 'On the Dangers of Stochastic Parrots' (2021)
- Wrote 'Voices in the Code' (2022), a study of the people and values behind the U.S. kidney allocation algorithm
- Built OpenAI's policy planning team in 2023, starting with negotiating its White House voluntary commitments on AI
- Led the writing of OpenAI's system cards, then resigned in 2026 with the Atlantic essay 'I Quit OpenAI Because Its Culture Is Broken'
Videos & Interviews
Connections
Paul Christiano
Influenced byFounder, Alignment Research Center · OpenAI Foundation Board
Christiano's statement on joining OpenAI's board in September 2026, that there is 'a meaningful risk' of 'catastrophic and irreversible loss of control in the very near term', is what Robinson names as the moment he looked around the company and asked what his loved ones, and strangers, would want OpenAI to be doing if Paul were right. He says he does not know whether Christiano is right. What moved him was that the caution such a warning would call for was 'totally unlike anything' he had seen in the industry. He resigned weeks later.
youtube.com · paulfchristiano.substack.com
Margaret Hamilton
KindredComputer Scientist & Software Engineering Pioneer
Hamilton called the usual way of building software an 'after the fact life cycle', where you write it and then hunt the errors, 'like back in the old medical community where they didn't wash their hands.' Half a century later Robinson left OpenAI describing the same pattern in AI, 'shipping new capability and risk every Tuesday,' and asked that a release wait until its safety criteria are met, however long that takes. He also pointed to the Challenger launch, where the risk 'had been known and documented and accepted.' One pairing, read across fifty years: safety is a property you design in, not one you discover after release.
youtube.com · youtube.com · youtube.com
