Fired OpenAI Safety Researchers Dispute Dismissals in Open Letter

Jasmine Wang, Mikita Balesni and Tomek Korbak posted an open letter to OpenAI arguing that the "very public manner" of the company's communications about their firings could have a "chilling" effect on employees at a time when "the world's safety depends on them."

OpenAI dismissed the three employees last week, saying they shared information with an external AI safety organization. In a statement, the company said the employees "violat[ed] our policies on accessing and handling sensitive company information... breaking the trust essential to our work." According to Engadget, the firings raised awkward questions because they came amid recent serious and potentially illegal OpenAI agent hacks of Hugging Face and other organizations, and they caught the attention of lawmakers.

"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 letter states. The researchers wrote that they could previously "raise safety concerns and disagree openly" and were encouraged to draw on the expertise of independent safety organizations, which they described as part of what made OpenAI special and why they were proud to have been part of the team.

The researchers said they "acted in line with OpenAI's mission and within the working norms of the time," and that the firing "leaves us worried that the norms inside OpenAI are shifting and that employees are now unclear on where they stand."

Addressing other accounts of the episode, Wang, Balesni and Korbak said they were not the leak source for a The Information article involving OpenAI's new, less monitorable architectures, and that they do not believe they engaged with external parties outside their mandate. Communication with external parties, they said, was done "in coordination and discussion with board members and the C-suite." Wang notified an executive after she accidentally clicked on a sensitive email, they added.

The letter made several recommendations to OpenAI: that the company adhere to its public commitments to embed third-party safety auditors and not use the firings as a "pretext for stepping away from those partnerships"; that it preserve the monitorability of frontier models; and that it 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 wrote that OpenAI leadership "is saying that they strongly agree with our letter," while voicing concern about the company's openness. "We were not the first to be pushed out of OpenAI under suspicious circumstances," she wrote. "Unless employees take a stand now against this kind of maneuver, I am concerned we will not be the last." She said the message to staff still at OpenAI was that raising concerns or working closely with outside safety groups could make them "next, without being told why."