The basic facts of the incident are not complicated: OpenAI announced last week that it had fired three employees, saying an internal investigation found they violated company rules on accessing and handling sensitive information and were suspected of sharing confidential information with third-party AI safety evaluation organizations outside established procedures. The three fired employees were Jasmine Wang, who worked on model alignment; Tomek Korbak, who worked on AI safety research and had participated in technical communication between OpenAI and external safety evaluation organizations; and Mikita Balesni, who worked on AI safety and model alignment research.
But the three then published an open letter on social media addressed to OpenAI leadership (the Safety and Security Committee, the Safety Advisory Group and the Mission Advisory Committee), offering a completely different explanation. Mikita Balesni said they were fired because they "put AI safety above OpenAI's short-term interests as a company." This statement is the individuals' claim and directly conflicts with the company's stated information-rule reason.
It must be made clear that neither side's account is currently supported by an independent third-party conclusion. OpenAI's stated reason for the firings and the three researchers' account each tell a different story; the open letter's content about personal backgrounds, the number of petition signatures and Altman's commitment are all the individuals' own accounts and lack first-party documentary support. This article is also based only on a single report by Machine Heart and cannot independently verify the full text of the open letter or OpenAI's official response.