Former OpenAI Safety Employee Resigns, Says AI Industry Culture Is ‘Broken’

Robinson’s criticism focuses on how frontier AI companies operate. He argues that the problem is deeper than adding a few new rules or regulations about how training models are handled. Silicon Valley, he says, has operated with “extreme confidence” and “perpetual sprints,” building bigger and better models with an “unimpeded optimism” that ignores or underestimates potential problems.

He says AI companies now need to develop a sense of humility and look outside the insular, move-fast-and-break-things world of the technology industry. In his view, AI development requires safeguards comparable to those used in other high-risk industries. “Given today’s risks, frontier labs need to run like nuclear power plants or busy airports, with layers of redundancy and careful, time-consuming planning, so that the occasional and inevitable human error does not open a door to disaster,” Robinson wrote.

The Verge noted that it is understandable for observers to feel cynical when people who helped build AI systems later warn about their dangers. But the publication added that this does not mean their warnings should be discounted.

Robinson is the latest in a series of researchers and safety workers who have left positions at prominent AI firms. According to The Verge, Jacob Coxon appears to have started the wave by quitting Anthropic and then publicly saying AI “could kill us all by the end of the decade.” He was followed by Robert O’Callahan, Bilal Chughtai and Josh Engels at Google DeepMind, as well as Joe Benton at Anthropic.

At OpenAI, Robinson’s former role placed him close to the process of documenting safety considerations for major model releases. His resignation and Atlantic editorial add another voice to internal and external debates over whether fast-moving AI labs can adequately manage the risks of the systems they produce. The Verge reported his comments and the broader pattern of departures.