Another OpenAI Safety Expert Quits and Raises New AI Safety Concerns
Another OpenAI Safety Expert Quits and Raises New AI Safety Concerns
October 05, 2026 
OpenAI safety veteran David Robinson resigns, warning that the company’s culture and fast AI development model could create bigger risks.
OpenAI safety veteran David Robinson knows how his resignation looks. He starts his essay by calling himself “something of a cliché”: an AI company employee who quits and then raises concerns about the company.
But Robinson was not a junior employee; he helped write the safety reports released with OpenAI’s major product launches and, after three and a half years at the company, says he was one of its longest-serving employees. He is leaving because he believes the company’s culture is broken.
“I resigned this week from OpenAI. I led the writing of the safety reports we published with each major launch. Now I’m joining a parade of former colleagues—at OpenAI and the industry’s other leaders—who have decided that the current path is unacceptable.” Robinson told The Atlantic. “I agree with other recently departed staff that the companies building this technology aren’t being nearly careful enough. But I believe that we need to look deeper than specific rules or new laws. We need to talk about culture.”
Robinson’s resignation echoes something Jacob Coxon said after leaving both OpenAI and Anthropic, when he accused these companies of gambling with people’s lives. That comment set off a real debate, pushing Anthropic’s CEO Dario Amodei to unveil a more cautious development plan and AI executives to sign a rushed, non-binding safety pledge after meeting with President Trump this week. Robinson thinks that kind of response misses the point entirely.
His argument isn’t about one bad CEO or one company’s specific failures. Most of the reporting on OpenAI has centered on how Sam Altman lost the trust of former colleagues, but Robinson’s essay argues the real problem is Silicon Valley’s culture writ large. New rules and new laws won’t fix a mindset problem.
Robinson says OpenAI relies on a trial-and-error approach. The company calls it iterative deployment: release systems, find problems, and then add or improve safeguards.
His concern is that this approach will inevitably lead to failures, and those failures could become more serious as AI systems become more powerful. He points to the recent breach of Hugging Face systems involving OpenAI agents, along with other rogue-agent incidents, as signs that this problem is already happening.
His comparison is blunt: frontier AI labs should be run like nuclear plants or major airports, with redundancy built in and planning that takes real time, so one human mistake doesn’t open the door to disaster. Then he adds the detail that actually stings. In his time at OpenAI he says he never worked alongside anyone with real experience keeping airplanes safe, keeping reactors from melting down, or keeping the financial system from collapsing.
OpenAI says it is already working on these issues, according to spokesperson Drew Pusateri, reported TechCrunch. The company says it is improving security in research environments, training models to act responsibly, increasing outside testing, and using stronger real-time monitoring to detect harmful behavior earlier. Robinson’s criticism is that this is still the same cycle of finding problems and fixing them after they appear.
Robinson says the problem goes beyond safety engineering and company culture. He argues that current methods for checking whether AI systems follow human values are still too basic for such a complex problem. The industry, he says, has no clear way to measure alignment, and that becomes more worrying as AI systems become more capable. If AI keeps improving while our ability to measure and control its behavior does not, the risks could increase.
Robinson is also open about how his resignation may look. He has hired a PR firm, something that has become common among AI whistleblowers, but says the decision to speak out was entirely his own.
He also says he could have stayed at OpenAI and tried to push for change from inside the company. But he felt that he and his colleagues were moving too fast to stop and think about the bigger changes the company needed.
Follow me on Twitter: @securityaffairs and Facebook and Mastodon
(SecurityAffairs – hacking, OpenAI)
About Author
