OpenAI Safety Researchers Challenge Firings, Raise Concerns Over Company Culture

OpenAI is facing questions over its handling of AI safety after three researchers — Jasmine Wang, Tomek Korbak and Mikita Balesni — were fired last week. The researchers have now written an open letter disputing the company’s explanation for their dismissals and warning that the move could discourage employees from raising safety concerns.

In their letter, the former employees stressed that as AI models become more advanced, companies must retain the ability to monitor how these systems operate. They specifically highlighted the importance of tracking a model’s chain-of-thought, which refers to the step-by-step reasoning process behind its responses.

Sharing the letter on X, Balesni alleged that their commitment to safety had put them at odds with the company’s immediate priorities. “I believe we were fired for prioritising safety over the near-term interests of OpenAI as a corporation,” he wrote.

Wang also reshared the post, claiming that other employees had experienced similar departures. “We were not the first to be pushed out of OpenAI under suspicious circumstances,” she wrote.

The researchers warned that the manner in which their employment was terminated could affect the company’s internal 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,” they said. “Terminations such as ours, executed and communicated so abruptly, are chilling the open culture OpenAI has prized in the past.”

Balesni said he was not given written reasons for his dismissal. According to him, he was told during an exit call that OpenAI no longer trusted him because he was “speaking too much to third party safety organisations”. He said he had never shared company intellectual property and that his work had been coordinated with “my reporting line, research leadership, and the board.”

Wang said she was dismissed over accessing an executive’s email account, which she claimed had been delegated to her for recruitment purposes. She added that she had asked the IT team to revoke the access once it was no longer required. “The reasons that we were provided for our terminations are simply not adding up,” she wrote.

Korbak said he was escorted out by security after being told the company no longer trusted him. He claimed his communication with external research group METR was cited as a reason for his dismissal. “To be clear, talking to METR was my job,” he wrote. He also said he had repeatedly warned about OpenAI “losing the ability to monitor what AI agents think.”

OpenAI has rejected the allegations, saying the three researchers were fired for violating policies governing sensitive company information, not for raising safety concerns. The company said an investigation uncovered a “pattern of misconduct” but did not publicly specify the alleged violations.

The researchers maintain that they acted in good faith, arguing that the dismissals could make employees hesitant to collaborate with outside experts or speak openly about AI safety.