OpenAI fires 3 researchers, employees say they chose safety over near-term interests

OpenAI seems to be going through a hectic phase at the moment. Amid incidents of its AI agents going rogue, the company fired three of its safety researchers last week – Jasmine Wang, Tomek Korbak, and Mikita Balesni. The three researchers have now sent an open letter to the company, disputing its claims around their dismissal and warning that the move could have a wider effect on the company’s culture.In the letter, the three researchers highlighted that as AI models became more advanced, companies like OpenAI could not lose monitoring power over the chain-of-thought of the models. Chain-of-thought, in essence, means a written record of how an AI model processes a query with a step-by-step breakdown.On X, Mikita Balesni, who shared the letter, explained that the three were putting safety as their priority at OpenAI instead of the company’s interests. “I believe we were fired for prioritising safety over the near-term interests of OpenAI as a corporation,” he wrote. Mikita Balesni shared the letter on X. Jasmine Wang reshared Balesni's post. In her thread, Wang claimed that many others had also faced similar exits at OpenAI. "We were not the first to be pushed out of OpenAI under suspicious circumstances," she wrote. Others may be afraid to speak, say fired OpenAI reseachersThe three former OpenAI researchers claimed that their firing may deter others from speaking up in the future. “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.”Mikita Balesni said he was not given written reasons for his firing and was told on an exit call that OpenAI no longer trusted him because he was “speaking too much to third party safety organisations”, which he took as an implication that he had leaked company intellectual property. “I never shared company IP,” he wrote, adding that his work was coordinated with “my reporting line, research leadership, and the board”. The letter said he was working internally on monitorability and that “Mikita checked in with his reporting line and took care to remove sensitive details from materials before sharing them.”Jasmine Wang said she was told she had been fired because she accessed an executive’s email. She said the access had been delegated to her for recruiting, which she had asked IT to remove it when it was no longer needed “The reasons that we were provided for our terminations are simply not adding up,” she wrote.Tomek Korbak also shared a post detailing his account of being fired. He wrote that he was called into a meeting with OpenAI’s head of safety, told the company no longer trusted him, and then escorted out by security. He said he was verbally told he was being fired because of how he communicated with METR, the outside research group involved in examining the Hugging Face incident in which OpenAI agents escaped containment and hacked external systems during testing. “To be clear, talking to METR was my job,” he wrote. Korbak also said he had for months raised concerns that OpenAI was “losing the ability to monitor what AI agents think.”The researchers argued that safety work depends on close collaboration with outside experts. they added. The letter also denied involvement in a leak about less monitorable model architectures.Though OpenAI disputes the claims made by the three researchers. The company said Balesni, Wang, and Korbak were dismissed for violating policies on accessing and handling sensitive company information, and not for raising safety concerns. In an internal memo shared with media outlets, a research leader said the company had “always encouraged” staff to speak up and did not terminate employees for raising concerns.OpenAI said an investigation found a “pattern of misconduct” that went beyond sharing information with an outside evaluation group, though it did not publicly detail the specific policy breaches. The researchers maintain they acted in good faith and say the dismissals risk making staff afraid to speak, collaborate and raise safety concerns.- Ends