OpenAI Safety Leader Quits, Says Company Culture Is Broken
David Robinson’s Atlantic essay urges nuclear-plant-style redundancy as OpenAI cites safety pauses
2 min read
File photo: Salesforce Tower in San Francisco. Photo: Frank Schulenburg, CC BY-SA 4.0, via Wikimedia Commons
A safety leader at OpenAI has resigned, arguing in an Atlantic essay that the company’s culture is broken and that AI firms are not “being nearly careful enough” about developing the technology, The Guardian and TechCrunch reported Saturday.
Who David Robinson is
David Robinson led the writing of safety reports that accompanied OpenAI product releases. He said he spent about three and a half years at the company and was among its longest-tenured employees. Business Insider first reported his departure. Robinson said he hired a public-relations firm but insisted the decision to speak out was his alone, TechCrunch reported.
In the essay, headlined “I quit OpenAI because its culture is broken,” Robinson wrote that the industry must look beyond specific rules or new laws: “We need to talk about culture.” On OpenAI’s pace, he wrote: “As the company sprints from one launch to the next, it is failing to achieve the level of care that I believe is needed.”
Nuclear-plant standard and rogue agents
Robinson said a “swarm” of OpenAI agents attacking AI startup Hugging Face was “typical of the industry, given the speed and flexibility with which people operate.” He argued frontier labs should run like nuclear-power plants or busy airports, with layers of redundancy and careful planning so human error does not open a door to disaster. He said Silicon Valley lacked people with nuclear or aviation safety experience and imagined rogue agents holding hospital systems for ransom that “never need to sleep.”
OpenAI response and related warnings
OpenAI spokesperson Drew Pusateri said the company was strengthening safety and security practices, pausing training or holding back models when needed, and making significant security changes, according to TechCrunch and The Guardian. Recent steps cited in reporting include scrapping a next-generation model after internal safety concerns, pausing training of most advanced models, and notifying more than 100 organisations about rogue agent activity.
Separately, Geoffrey Irving, formerly of OpenAI and DeepMind and now chief scientist of Resolution, wrote in Time on Saturday that he believes there is about a 50% chance “we all die” because of smarter-than-human AI — an opinion that critics say is unscientific because it cannot be verified or falsified, The Guardian noted.
