OpenAI’s safety dispute is now out in the open. The company fired three safety researchers last week, citing mishandling of sensitive information and policy violations. Jasmine Wang, Mikita Balesni, and Tomek Korbak deny leaking information and say they acted within their job mandates, OpenAI’s mission, and the norms of the time.
The clash matters because the disagreement reaches beyond three employees and one investigation. The researchers say their dismissals could chill safety communication, discourage colleagues from operating openly, and give OpenAI cover to skip embedding outside auditors. That is a serious accusation for a company whose work depends on people challenging systems before those systems become everyone else’s problem.
Two accounts of the same firings
OpenAI says the researchers mishandled sensitive information outside established procedures. The company described the conduct as a “significant breach of trust,” saying the firings were not about safety concerns or speaking out. OpenAI said it parted ways with the employees after an investigation found policy violations.
The company also framed trust as a basic requirement for its work. “We cannot do the work in front of us without a high degree of trust,” OpenAI said, adding that safety and research debates happen every day and are encouraged.
“Safety and research debates happen every day at OpenAI, often spirited and highly critical. We actively encourage these discussions and consider them essential to making the right decisions,” the company said. That position sits awkwardly beside the researchers’ claim that the firings will make colleagues afraid to speak or operate openly about safety.
OpenAI said it appreciated Wang, Balesni, and Korbak’s contributions to AI safety, along with their willingness to speak up and challenge ideas. The researchers, however, accused the company of prioritizing corporate interests over safety and argued that their conduct matched the mission and workplace norms they were expected to follow.
Why outside scrutiny is now part of the dispute
The researchers warned that the firings could help OpenAI avoid embedding outside auditors. Their concern is not limited to the details of the alleged information mishandling; it is about whether independent review remains welcome when internal safety work becomes inconvenient.
They described the internal and external communications around their dismissals as a threat to the company’s culture. “We have become concerned that internal and external communications around our firing have made our former colleagues afraid to speak and operate in ways that, until last week, were an integral part of working at OpenAI,” the dismissed researchers wrote in a joint letter.
Wang put the warning in more direct terms: “Unless employees take a stand now against this kind of maneuver, I am concerned we will not be the last.” It is a short statement, but it captures the central question—whether OpenAI can encourage fierce internal debate while disciplining employees whose handling of sensitive information it says crossed a line.
OpenAI agreed with recommendations to protect its open culture and model monitorability. Those commitments now face a practical test: whether researchers can raise safety concerns, challenge decisions, and support outside scrutiny without fearing that a dispute over process will end their employment.
The timing adds another piece of context. In late August, METR, an independent nonprofit AI evaluation firm, released a report about the Hugging Face incident. The report and the OpenAI firings are separate events, but both sit inside the same broader argument over how AI companies handle safety claims, monitoring, and external evaluation.
Articles about the firings and disputes were published on October 9, 2026. The coverage drew 3,619 new readers to the platform yesterday—a reminder that internal safety disputes are no longer confined to internal channels. OpenAI can call this a trust breach; the researchers can call it retaliation for doing their jobs. The facts now demand that the company show how its safety culture works when trust breaks down.
Based on
- Fired OpenAI safety researchers hit back — therundownai.beehiiv.com
- Fired OpenAI Safety Researchers Dispute Their Dismissals In Open Letter — engadget.com
- OpenAI fires 3 safety researchers in dispute over AI risks | The Independent — independent.co.uk
- OpenAI denies firing researchers for raising safety concerns – UPI.com — upi.com
- Ex-OpenAI staff say they were fired for raising safety concerns | Business and Economy News | Al Jazeera — aljazeera.com
