What We Know About OpenAI's Firing of Three Safety Researchers

Three former OpenAI researchers say the company fired them for raising safety concerns, a claim OpenAI disputes, pointing instead to what it calls mishandled sensitive information, according to the BBC.

The dispute surfaced in a public letter and a series of posts on X from the three former employees, who worked on safety or alignment at the company before their dismissal. OpenAI's response, issued twice within a week, says an internal investigation found conduct beyond what the researchers described in their own account.

What the fired researchers allege

"I believe we were fired for prioritising safety over the near-term interests of OpenAI as a corporation," Mikita Balesni wrote on X alongside an open letter to OpenAI leadership, the BBC reported. Balesni co-authored the letter with Tomek Korbak and Jasmine Wang, all described as former safety or alignment employees at the firm.

The three said their firings have made remaining staff "afraid to speak and operate in ways that, until last week, were an integral part of working at OpenAI," according to the letter cited by the BBC. Wang went further in a separate post, writing, "We were not the first to be pushed out of OpenAI under suspicious circumstances. Unless the employees take a stand now against this kind of manoeuvre, I am concerned we will not be the last."

What OpenAI says happened instead

OpenAI has twice rejected the researchers' framing. A company spokesperson told the BBC on October 2 that an internal investigation "confirmed that these individuals mishandled sensitive information outside established company procedures." Research leaders repeated and sharpened that position in a note issued Friday, October 9, the BBC reported.

"We want to be very clear that these decisions were not about raising safety concerns or speaking out," the note said, per the BBC. "Our internal investigation uncovered a significant breach of trust beyond what's outlined in the letter they published and we stand by the decision to not continue their employment."

OpenAI has not detailed what the alleged breach involved beyond the general description of mishandled sensitive information, and the company's statements do not identify specific documents, systems, or recipients tied to the alleged conduct.

Who are the three researchers

The letter names Mikita Balesni, Tomek Korbak and Jasmine Wang as its authors, each described in the BBC's reporting as a former safety or alignment employee at OpenAI. Alignment work, as the BBC explains it, involves building human ethical principles into AI systems so the technology's behavior stays consistent with what people value.

What specific safety concern did Korbak raise

Korbak tied his dismissal to a narrower technical worry about oversight of increasingly autonomous AI systems. "For months, I'd been raising safety concerns that we're losing the ability to monitor what AI agents think, one of our best tools for catching when they misbehave," he wrote on X, as reported by the BBC. "I believe that was why I was fired."

That concern touches on a method researchers use to review the internal reasoning AI systems generate before they act, a technique some in the field treat as an early-warning system for unwanted behavior. Korbak's post does not specify which product or research effort the monitoring concern applied to.

What happens next

OpenAI told the BBC it is finalizing contracts with third-party safety assessors and plans to announce details in the coming weeks, though the company has not said whether that effort is connected to the dismissals or to the broader debate over internal AI safety oversight. Neither side has released the underlying investigation findings or the full text of the breach allegations publicly.

The dispute lands amid wider scrutiny of how AI firms handle internal dissent over safety practices. The BBC noted its reporting follows other recent coverage of AI safety concerns, including research on teen use of ChatGPT and public comments from a former Anthropic researcher about risks the technology could pose.

Key dates in the dispute

• October 2, 2026: OpenAI spokesperson tells the BBC its investigation found the researchers mishandled sensitive information outside company procedures.
• October 9, 2026: OpenAI research leaders issue a follow-up note reaffirming the decision and alleging a breach of trust beyond what the letter describes.
• Three former employees — Balesni, Korbak and Wang — co-signed the public letter disputing OpenAI's account.

OpenAI has not responded publicly to Wang's claim that the three were not the first employees "pushed out" under disputed circumstances, and the BBC's report does not name other individuals in that category.