David Robinson, a safety researcher at OpenAI who led the writing of safety reports accompanying every major model release, has resigned from the company. In an essay published in The Atlantic, Robinson described OpenAI's culture as "broken" and explained why he felt compelled to speak out after three and a half years — making him among the longest-tenured employees at the company.
Robinson's departure is notable not because he was a senior executive, but because of his role: he was the person responsible for documenting the risks and limitations of OpenAI's most powerful models before they shipped. His safety reports were among the few internal documents that regularly reached external audiences, and they carried weight because they came from someone who had seen the models' capabilities up close.
By his own admission, Robinson is "something of a cliché": an employee at a leading AI company who issues a dire warning while resigning. But the specifics of his critique — and the fact that he stayed as long as he did — give his account more gravity than the typical departure post.
The resignation arrives at a sensitive moment for OpenAI. The company has been restructuring its safety team, and several other researchers have left in recent months. Critics have argued that OpenAI's shift toward a more commercial posture — including its push for a for-profit conversion — has come at the expense of the safety-first culture that distinguished it in its early years.
OpenAI has not yet responded to Robinson's essay. The company has previously stated that safety remains central to its mission, even as it scales its operations and restructures its organizational chart.
