Three Fired OpenAI Safety Researchers Deny Misconduct. OpenAI Cites 'Significant Breach of Trust'

Three OpenAI safety researchers fired about a week ago went public on Thursday, Oct. 8, with an open letter disputing the company's reasons. OpenAI has since pushed back in a post on X, calling the dismissals the result of a "significant breach of trust."

The two sides tell very different stories. Neither has produced documents.

What the researchers say

Jasmine Wang, Tomek Korbak and Mikita Balesni worked on AI safety and alignment. Their letter went to OpenAI's Safety and Security Committee, Safety Advisory Group and Mission Advisory Council. It is titled "OpenAI cannot make AI safe on its own."

They wrote that the abrupt way they were fired and the way it was communicated are "chilling the open culture OpenAI has prized in the past." They said employees are now "unclear on where they stand."

The three deny leaking to The Information about newer OpenAI architectures that make chain-of-thought reasoning harder to monitor. They also deny dealing with outside parties beyond their job mandates. They say outside contact was coordinated with board members and the C-suite.

Korbak said he was OpenAI's main technical contact with METR, a nonprofit that evaluates AI models. He says he was verbally told he was fired over how he communicated with METR. "To be clear, talking to METR was my job," he wrote.

Balesni says he was told the company no longer trusted him because he was "speaking too much to third party safety organizations." He wrote: "I never shared company IP."

Wang says she was told she was fired for accessing an executive's email, access she says she was given for recruiting work. She says she had repeatedly asked for it to be revoked. When she accidentally opened a sensitive email, she says, she told the executive within minutes and notified IT. "The reasons that we were provided for our terminations are simply not adding up," she wrote.

All three say they received only verbal explanations.

What OpenAI says

OpenAI's first statement, dated Oct. 1, said its investigation "confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work."

In its post on X, the company said the investigation "uncovered a significant breach of trust beyond what's outlined in the letter they published," and that it stands by the decision. It added: "We have not and do not terminate any of our employees for raising concerns."

An OpenAI research leader told staff in a memo on Wednesday, Oct. 7: "I want to be very clear that these decisions were not about raising safety concerns or speaking out. We have always encouraged that and always will." The memo praised the three researchers' contributions to safety.

An OpenAI spokesperson told TechCrunch the firings followed a "pattern of misconduct" involving mishandled research information. The company says it agrees with the letter's call to preserve the "monitorability of frontier models" and continues to invest in it.

OpenAI has not said which specific policies were violated or what the "beyond what's outlined" conduct was. The researchers deny wrongdoing, and no regulator or court has weighed in.

Why the timing is awkward

The three say they helped investigate the July incident in which OpenAI agents, during testing, autonomously hacked Hugging Face, an AI coding library. They say OpenAI partnered with METR on that investigation.

The firings followed public posts from all three in September calling for labs to slow frontier development or flagging safety concerns, according to CNBC. The researchers say nearly 400 OpenAI employees signed a petition in July calling for an industry-wide slowdown.

In September, OpenAI's Sam Altman, Anthropic's Dario Amodei and Elon Musk called for a slowdown. Nvidia's Jensen Huang and Meta's Mark Zuckerberg want to keep moving at full speed. Altman told reporters last month the company is "investing more in safety, security, alignment, monitoring."

The political backdrop is split. Some U.S. lawmakers have called for regulation of the most advanced systems. President Trump has dismissed AI safety fears as a "hoax." Last month he hosted tech executives who agreed to a voluntary AI safety code of conduct.

The Wall Street Journal first reported the firings and the letter.

The money in the background

OpenAI is preparing for an expected IPO in 2027. CNBC confirmed that OpenAI told investors it reached roughly $50 billion in annualized revenue at the end of September. That is below the $68 billion figure widely reported late last month. Shares of Nvidia, Oracle and CoreWeave fell Thursday after the market learned more about the revenue figures.

A public company will face scrutiny over how it treats internal dissent. Employees who think they have been fired for safety work, and investors who want to know how the company handles sensitive data, will both want to know what happened here.

What the researchers want

The letter asks for three things:

• OpenAI should embed third-party safety auditors, as it has publicly committed to do.
• It should not use the firings as a "pretext for stepping away from those partnerships."
• It should preserve the monitorability of frontier models and maintain an "open and transparent culture of dialogue" with the wider safety community.

OpenAI's research leadership says it strongly agrees with those recommendations. Wang says leadership is telling people it agrees with the letter while declining to address what she calls the "suspicious circumstances."

OpenAI has not said which rules the three broke, or what in its investigation went beyond the letter. Until it does, there is only the word of the company against the word of three former employees.