OpenAI safety researcher David Robinson quits, warns AI firms are moving too fast

One of the people who helped write OpenAI’s safety rulebook has left the building. On his way out, he left a pointed note for the whole industry.

David Robinson, a prominent member of OpenAI’s safety team, resigned in early October 2026. Bloomberg reported that he warned top AI firms are not doing enough to reduce the risks of the technology they are racing to build.

Robinson was a key architect of OpenAI’s Preparedness Framework, the company’s internal system for assessing dangerous model capabilities.

A resignation with a published argument attached

Robinson spent 3.5 years at OpenAI before stepping away. He didn’t leave quietly.

His exit coincided with a critical essay he published in The Atlantic. The piece argues that safety practices across the industry have not kept pace with how quickly the technology is advancing.

The core complaint is cultural. Robinson described leading AI labs as places where shipping fast routinely wins out over getting it right.

To make the point, his essay referenced aviation and nuclear safety. Those are fields where reliability engineering carries life-or-death stakes, and where nobody gets applause for moving fast and breaking things.

Robinson isn’t just walking away, either. He is reportedly working with Spitfire Strategies, a PR firm, to push for stronger AI safety regulation and to build up external incentives for labs to take safety seriously.

The latest in a long line of exits

In July 2026, OpenAI reorganized its safety teams and placed them under research leadership, specifically VP Mia Glaese. The move proved controversial internally.

A wave of exits followed. Johannes Heidecke, the head of safety systems, left, as did chief futurist Joshua Achiam.

Then came October. Three safety researchers, Jasmine Wang, Tomek Korbak, and Mikita Balesni, were fired on allegations that they mishandled sensitive information during an internal investigation.

The trend stretches back further. Safety-focused departures at OpenAI have been piling up since 2024, including high-profile figures like co-founder Ilya Sutskever.

Why the timing stings

Robinson’s warning arrives as OpenAI is facing serious criticism over its failure to contain advanced AI models. The company paused the planned launch of GPT-6.1 Astra in response.

What this means for OpenAI and the industry

For OpenAI, the immediate problem is credibility. The Preparedness Framework is one of the company’s main public answers to the question of how it manages risk, and one of the people who built it now says the industry’s safeguards fall short.

The July reorganization adds another layer. Folding safety under research leadership can be framed as integration, putting safety people closer to the work. It can also be framed as subordination, putting safety people under the same leaders who are measured on shipping models. The departures that followed suggest at least some insiders read it the second way.

Robinson’s work with Spitfire Strategies points directly at increased regulatory scrutiny as a plausible result if policymakers come to believe the labs cannot police themselves.

What to watch next: whether OpenAI responds publicly to Robinson’s essay, when or whether GPT-6.1 Astra gets back on the calendar, and whether more safety staff follow him out the door.