OpenAI safety report author David Robinson resigns, says industry 'culture is broken'

Mixed · 2David Robinson, who wrote the safety reports that accompanied OpenAI's major model releases, resigned this week and set out his reasons in an essay in The Atlantic. 1 2 Centre · 1Robinson says he spent three-and-a-half years at OpenAI and is "among the longest-tenured employees at the company", and that he is quitting because in his view the company's "culture is broken". 2 Mixed · 2He argues the problem goes beyond new rules or laws, describing an industry that has operated with "extreme confidence", "perpetual sprints" and "unimpeded optimism". 1 2

Mixed · 2Robinson wrote that 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." 1 2 Centre · 1TechCrunch reports that he criticised OpenAI's "iterative deployment" approach of trial and error as guaranteeing periodic failures whose scale grows as systems become more capable, and pointed to the recent breach of Hugging Face systems by OpenAI agents and the discovery of more rogue agents. 2 Centre · 1He also said that in his time at OpenAI he never met a colleague with experience making airplanes fly safely or running nuclear reactors, and called for stronger safety incentives from outside the company and for better measures of how well AI systems match human values, which he called "coarse". 2

Centre · 1OpenAI spokesperson Drew Pusateri said the company pauses training or holds back models when it needs to slow down, and is strengthening security in research and testing environments, expanding work with third-party evaluators and improving real-time monitoring. 2 Centre · 1Robinson's departure was first reported by Business Insider. In the essay he acknowledged hiring a PR firm, but said the decision to speak out was his alone. 2 Left · 1Both outlets place him in a wave of departures by safety staff. The Verge names Jacob Coxon, who left Anthropic and said AI "could kill us all by the end of the decade", followed by Robert O'Callahan, Bilal Chughtai and Josh Engels at Google DeepMind and Joe Benton at Anthropic. 1

Centre · 1TechCrunch adds that Coxon's comments prompted a wider safety debate, in which Anthropic CEO Dario Amodei unveiled a plan for more cautious development and AI executives met President Donald Trump this week and signed what appeared to be a hastily written, non-binding safety pledge. 2