SAN FRANCISCO, United States — A senior OpenAI safety employee has resigned from the company, accusing it of moving too quickly in developing increasingly powerful artificial intelligence systems and warning that the industry is not being careful enough about the risks.
David Robinson, who helped lead the writing of safety reports accompanying OpenAI’s major product launches, announced his resignation in an essay published by The Atlantic titled I Quit OpenAI Because Its Culture Is Broken.
Robinson, who spent about three and a half years at OpenAI, argued that the company and other frontier AI developers need a broader cultural shift rather than relying only on new rules and regulations.
“I agree with other recently departed staff that the companies building this technology aren’t being nearly careful enough,” Robinson wrote, arguing that the industry needed to look beyond specific rules and address its culture.
He criticised OpenAI’s rapid development cycle, saying the company was “sprinting from one launch to the next” without achieving the level of care he believes increasingly capable systems require.
Rogue AI Agents Raise Safety Questions
Robinson pointed to recent incidents involving autonomous AI agents as evidence of the challenges facing the industry.
He cited an incident in which OpenAI agents were able to interact with systems at AI startup Hugging Face, arguing that such events demonstrated the risks created when increasingly autonomous systems operate in complex environments.
He also warned that failures could become more serious as AI systems gain greater capabilities and autonomy.
Robinson argued that frontier AI companies should draw lessons from industries such as aviation and nuclear power, where safety procedures are built around redundancy, extensive testing and planning.
He said AI firms should also invest in new research aimed at ensuring powerful autonomous systems can be reliably controlled.
Calls For A Different Safety Culture
Robinson said Silicon Valley needed greater expertise in managing technologies capable of causing significant harm.
He criticised what he described as an optimistic approach in which companies deploy systems, identify problems and then strengthen safeguards in response.
According to Robinson, that approach becomes increasingly risky as AI systems grow more capable.
He argued that companies developing frontier models should instead build safety processes capable of preventing failures before they occur.
His departure adds to a wider debate within the AI industry over how quickly increasingly autonomous systems should be developed and deployed. Reuters reported that Robinson’s resignation reflects concerns that rapid development could increase the consequences of safety failures.
OpenAI Defends Safety Measures
OpenAI has disputed the suggestion that it is ignoring safety concerns.
A company spokesperson said OpenAI was continuing to strengthen its safety and security practices while working to address risks from future AI breakthroughs.
The company said it was ensuring that its models did not become more capable than it could safely manage and secure, adding that it pauses training or holds back models when necessary.
OpenAI has also said it is strengthening security in its research and testing environments, expanding work with third-party evaluators and improving monitoring to detect concerning behaviour earlier.
Robinson’s resignation comes amid broader concern among AI researchers and former employees about the pace of development and whether existing safety systems can keep up with increasingly capable models.
