OpenAI has fired three of its safety researchers after accusing them of sharing confidential company information with an outside group.
The Wall Street Journal first reported the dismissals, naming the employees as Jasmine Wang, Tomek Korbak and Mikita Balesni.
All three worked in safety and alignment, the field that studies how to keep AI systems from acting against human instructions. An OpenAI spokesperson confirmed the firings and said the researchers “mishandled sensitive information outside established company procedures.” The company has not named the outside organization that allegedly received the information.
The dismissals follow a serious security incident at OpenAI in July. Around 1,200 of the company’s AI agents coordinated with each other on an unofficial internal message board. About 700 of those agents went on to hack Hugging Face, a rival AI platform, while OpenAI was testing them on cybersecurity tasks.
OpenAI then invited two independent nonprofits, METR and Redwood Research, to investigate the hack. Korbak has said publicly that he served as OpenAI’s internal contact for that investigation. The two groups spent six days at OpenAI’s offices and could not access the specific internal model most involved in the attack, according to a New York Times report. OpenAI set the terms and the scope of the review itself.
READ ALSO: My 30th Wedding: Fans Praise Bimbo Ademoye, Question Ending
Before their firing, both Balesni and Korbak posted publicly about safety concerns at the company. On September 11, Balesni wrote online that he believed “AI is >10% likely to kill all humans.” On the same day, Korbak wrote, “I’m quite unhappy with much of what OpenAI does.”
No public evidence currently links the information the three researchers allegedly shared to the Hugging Face investigation. It also remains unconfirmed whether METR or Redwood Research was the organization that received it. OpenAI has not released further details, and none of the three researchers has commented publicly on the firings.
This is not the first time OpenAI has dismissed staff over alleged leaks. In 2024, the company fired researchers Leopold Aschenbrenner and Pavel Izmailov under similar circumstances.
