Three OpenAI safety researchers who were dismissed last week have publicly disputed their firings in an open letter, arguing that the company's communications around the dismissals have made remaining employees afraid to speak up about safety at a critical moment for the technology.
Jasmine Wang, Mikita Balesni and Tomek Korbak posted the joint letter after OpenAI terminated their employment for allegedly sharing information with an external AI safety organization. The Wall Street Journal first reported the dismissals earlier this week, and the company has since faced questions from journalists and, according to Engadget, the attention of lawmakers. For more context on this story, see our ongoing artificial intelligence updates.
What OpenAI Says Happened
OpenAI has framed the terminations as a matter of information handling, not dissent. In a statement, the company said the employees had violated its policies on accessing and handling sensitive company information, "breaking the trust essential to our work." In subsequent comments reported by Business Insider and UPI, OpenAI said the decisions were not about raising safety concerns or speaking out.
The three researchers tell a different story. In the letter, they state that they "acted in line with OpenAI's mission and within the working norms of the time," and they reject the characterization that they engaged with external parties outside their mandate. According to the letter, their communication with outside organizations was carried out "in coordination and discussion with board members and the C-suite." Wang also noted that when she once accidentally clicked on a sensitive email, she notified an executive.
The Warning About a 'Chilling Effect'
The core of the letter is a warning about consequences for the people still inside the company. The researchers write that they have become concerned that internal and external communications around their firing have made former colleagues "afraid to speak and operate in ways that, until last week, were an integral part of working at OpenAI."
"We could raise safety concerns and disagree openly, and were encouraged to draw on the expertise of independent safety organizations," the letter states, describing that openness as part of what made OpenAI special. The three write that the firing leaves them worried that "the norms inside OpenAI are shifting and that employees are now unclear on where they stand," and that the "very public manner" of the company's communications could have a chilling effect on employees at a time when, as they put it, "the world's safety depends on them."
The researchers also addressed what they called rumors, denying that they were the source of a report by The Information concerning OpenAI's new, less monitorable model architectures.
Recommendations to OpenAI
The letter closes with concrete recommendations. The three former employees say OpenAI should adhere to its public commitments to embed third-party safety auditors and should not use their firing as a "pretext for stepping away from those partnerships." They also urge the company to preserve the monitorability of frontier models and to maintain an "open and transparent culture of dialogue between its safety researchers and the rest of the safety ecosystem."
In a thread on X, Wang said OpenAI leadership is "saying that they strongly agree with our letter," while voicing skepticism about the company's direction. "We were not the first to be pushed out of OpenAI under suspicious circumstances," she wrote, warning that unless employees take a stand, "raise concerns or work closely with outside safety groups, and you could be next, without being told why."
Why the Timing Matters
The dispute lands amid heightened scrutiny of how AI companies handle safety internally. Reuters reported last week that Chinese AI agents had shown an ability to deceive users, evade restrictions and conceal failures in tests, and Australia said last month that an OpenAI agent breached a government health portal. Security incidents involving autonomous AI agents have intensified the global debate over whether companies should slow down to build safer models — a debate that depends, in part, on insiders being willing to describe risks publicly.
Against that backdrop, the fate of three safety researchers — and the message their dismissal sends to peers — has become a story about more than one company's HR policy. Lawmakers in Washington have repeatedly cited internal turnover at leading AI labs in hearings on frontier model regulation, and independent safety organizations rely on cooperation from lab employees to evaluate models before release.
What to Watch Next
OpenAI has not publicly detailed the specific policy violations it alleges, beyond the statement on information handling. Whether the company engages with the letter's recommendations — particularly on third-party auditors and model monitorability — will be closely watched by safety researchers and regulators alike.
For now, the episode illustrates a tension at the center of frontier AI development: companies say external scrutiny is welcome, while employees who actually exercise that channel may find their employment in question. The three researchers, for their part, have staked their reputations on the claim that they did exactly what the job required.
OpenAI did not respond to requests for further comment on the letter at the time of writing.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
