OpenAI safety employee resigns, warns company culture is broken

David Robinson, who wrote the safety reports accompanying OpenAI’s major model releases, has resigned from the company, The Verge reports. He is now speaking publicly about his departure in an editorial published in The Atlantic.

According to the available summary, Robinson’s remarks concern the culture inside the industry, though the source material provided does not complete that characterization or specify further detail.

The Verge’s report notes that Robinson is one of several people associated with AI safety work who have left major companies and subsequently spoken out about concerns tied to their former employers. The piece acknowledges that such warnings can invite skepticism, given that those raising them were themselves involved in building the systems they now caution about. The Verge argues that this history does not, on its own, justify dismissing what they say.

The report identifies Robinson specifically as a former author of safety documentation tied to OpenAI’s model releases, a role that placed him inside the company’s internal review process for new systems before they reached the public.

Further detail on the content of Robinson’s editorial, including any specific claims he makes about internal practices at OpenAI, is not included in the source material available at this time. The Verge’s full report and Robinson’s Atlantic editorial may contain additional specifics beyond what has been summarized here.