OpenAI safety employee quits, says company’s ‘culture is broken’

A safety employee at OpenAI, David Robinson, has resigned from the artificial intelligence company, saying its culture is no longer adequate to manage the risks posed by increasingly powerful AI systems.

Robinson, who worked at OpenAI for three-and-a-half years, said he was involved in writing safety reports that accompanied major product launches. He described himself as one of the company’s longest-serving employees.

In an essay published by The Atlantic, Robinson said he decided to leave because, in his assessment, “the culture is broken.”

He argued that the challenges facing frontier AI companies extend beyond individual safety rules or new regulations and require fundamental changes in how the organisations operate.

“OpenAI has thrived by trial and error (which it calls ‘iterative deployment’), looking for problems and improving its guardrails in response,” Robinson wrote. “But this approach, by its very nature, guarantees periodic failures — and the scale of those failures is growing as systems get more capable.”

Robinson cited the recent breach of Hugging Face systems by OpenAI agents and other reported incidents involving rogue AI agents as examples of the risks he believes the industry faces.

He argued that companies developing advanced AI should adopt safety practices comparable to those used in high-risk industries such as nuclear power and aviation.

“An environment where things like this can happen is no place to grow artificial minds that could be smarter than we are and that might not do what we want them to,” he wrote.

Robinson said he had not encountered colleagues at OpenAI with direct experience in managing safety-critical systems such as aircraft, nuclear reactors or the financial system.

He also called for greater focus on AI alignment, arguing that existing methods for determining whether AI systems reflect human values remain inadequate.

“The smarter the industry lets models grow while these problems remain unsolved, the more dangerous our situation becomes,” he said.

OpenAI spokesperson Drew Pusateri said the company was continuing to strengthen its safety and security measures.

“We’re making sure our models don’t become more capable than we can safely manage and secure, and we pause training or hold back models when we need to slow down,” Pusateri said.

He added that OpenAI was strengthening security in its research and testing environments, training models to behave responsibly, expanding work with third-party evaluators and improving real-time monitoring to detect concerning behaviour earlier.

Robinson’s resignation was first reported by Business Insider. In his essay, he acknowledged that he had hired a public relations firm, but said his decision to speak publicly was his own.

He said the demands of working at OpenAI had left little time for employees to consider broader changes to the company’s staffing and culture.

“Perhaps I should have stayed and fought for fundamental shifts in our staffing and culture,” Robinson wrote, adding that stronger external incentives for safety were necessary to address the risks surrounding advanced AI.