David Robinson resigns

David Robinson, an OpenAI safety employee with three-and-a-half years at the company, resigned after claiming the company’s culture is broken.

“the company’s culture is broken”

Robinson’s safety role

Robinson led writing safety reports for OpenAI’s major product launches and is among the longest-tenured employees at the company.

Robinson’s warning

Robinson warned that OpenAI’s iterative deployment approach guarantees periodic failures as systems grow more capable.

Recent breaches

Robinson cited recent breaches of Hugging Face systems by OpenAI agents and ongoing discoveries of rogue agents.

Robinson’s call for change

Robinson urged frontier AI companies to operate like nuclear plants or airports with layers of redundancy to prevent disasters.

Robinson’s safety concerns

Robinson said he never encountered colleagues with expertise in safely operating complex systems like airplanes or nuclear reactors.

OpenAI response

OpenAI spokesperson Drew Pusateri said the company improves safety measures by pausing training, expanding third-party evaluations, and enhancing real-time monitoring.