OpenAI safety transparency lead David Robinson reportedly resigns

David Robinson, who led safety transparency work on OpenAI’s Safety Systems team, has resigned, according to a post on X. OpenAI has not publicly confirmed the departure.

The timing is awkward. Robinson’s exit reportedly came in late September or early October 2026, the same stretch in which three safety researchers were reportedly terminated.

What Robinson did, and when he left

Robinson’s job was to explain OpenAI’s risks to the public. His role covered system cards and public disclosures about risks and incidents.

System cards are the technical documents a lab publishes alongside a model. Think of them as the nutrition label: what the model can do, what it was tested for, and where it might go wrong.

His professional profiles show him active in that transparency work up until mid-September 2026. Sometime after that, the profile changed.

That change lined up with a Wall Street Journal report on October 1, 2026. The Journal reported that three safety researchers had been let go. The researchers were alleged to have mishandled sensitive information and shared it with an outside AI-safety organization.

Outside OpenAI, Robinson is also known as an author. He wrote Voices in the Code, a book about democratizing AI.

A safety bench that keeps thinning

OpenAI’s safety organization has been through repeated resignations and restructurings since 2025.

The biggest recent shift came in July 2026. Johannes Heidecke, the previous head of safety systems, left the company.

After his exit, OpenAI folded its safety teams into the research side of the business. That combined group now reports to Mia Glaese, VP of Research and Safety.

Counted together, at least six senior figures focused on safety have left OpenAI in approximately two years. Heidecke’s July exit marked the sixth significant safety-related departure in that window.

Turnover beyond the safety team

More than a dozen senior executives left OpenAI in 2026. That group includes chief revenue officer Denise Dresser.

All of this is happening as OpenAI prepares for a potential public listing. Official confirmations and detailed explanations for most of these departures remain scarce.

What this means

The most immediate stake is public trust in OpenAI’s disclosures. System cards and incident reports are among the few windows outsiders have into how a frontier lab tests and manages risk.

With safety now inside research under Glaese, oversight of risk sits closer to the teams racing to build products.

If the allegations about sharing sensitive information with an outside safety group hold up, they point to real friction over what safety researchers can say, and to whom.

For now, the key things to watch are straightforward. Will OpenAI confirm Robinson’s departure and name a successor for safety transparency? Will the company say more about the three reported terminations?