OpenAI safety lead David Robinson resigns over company culture

David Robinson, who led the writing of the safety reports published alongside OpenAI’s major product launches, resigned this week after three and a half years at the company. He said he was one of its longest-serving employees and that he left because he believes its culture is broken.
Writing in The Atlantic, Robinson argued that the industry’s problems run deeper than specific rules or new laws. He criticised OpenAI’s iterative deployment model, in which systems are released, problems are found and safeguards are added afterwards, and said frontier labs should be run like nuclear plants or major airports, with redundancy and slow, deliberate planning.
He pointed to a recent breach of Hugging Face systems involving OpenAI agents as evidence that the approach is already producing failures, and said he never worked alongside anyone with experience of keeping aircraft safe or preventing reactor meltdowns. He also said the industry has no clear way to measure alignment.
OpenAI spokesperson Drew Pusateri said the company is improving security in research environments, training models to act responsibly, increasing outside testing and using stronger real-time monitoring. Robinson said that response remains the same find-and-fix cycle.