OpenAI fired safety researchers Tomek Korbak, Jasmine Wang and Mikita Balesni in the week before October 8, 2026. OpenAI said an investigation found policy violations involving sensitive information and denied the dismissals were about raising safety concerns. The researchers described different circumstances connected to their safety work and communications with outside organizations.
OpenAI’s explanation for the dismissals
OpenAI said its investigation found that the researchers violated policies on accessing and handling sensitive company information. The company also denied that the dismissals were about safety concerns or speaking out.
The company’s explanation is one side of a dispute over the three decisions. The researchers’ accounts point to their work on AI safety and contact with outside organizations, but they differ in the circumstances each person described.
What Korbak, Wang and Balesni said
Tomek Korbak said he was told that his communications with METR were the reason for his dismissal. He also said that communicating with METR, an independent nonprofit AI evaluation organization, was part of his job.
Jasmine Wang said she had access to an executive’s email for recruiting work and had repeatedly asked for that access to be revoked. She said she reported accidentally opening a confidential email.
Mikita Balesni said he was told he had communicated too much with outside safety organizations. He described his work as involving cross-company efforts on AI monitoring and denied leaking OpenAI intellectual property.
The researchers’ request for outside monitoring
In a public letter, the three researchers urged OpenAI to retain third-party safety monitors and preserve the ability to monitor frontier AI models—the most advanced systems being developed. They warned that the dismissals could discourage open discussion of AI safety concerns.
