OpenAI fires three researchers over handling of sensitive data

OpenAI confirmed the dismissal of three researchers after an internal investigation determined they mishandled sensitive information outside established procedures. The San Francisco-based company stated to AFP that the actions violated corporate policies and broke the trust required for its work.

OpenAI did not officially identify the three employees. According to The Wall Street Journal and Bloomberg, at least two of the individuals worked in artificial intelligence safety and alignment. Breaches identified by the company included sharing information with an outside group focused on AI evaluation, which OpenAI did not publicly name.

The Wall Street Journal reported that the dismissed staff had publicly raised concerns regarding AI risks and supported slowing development. One researcher, Tomek Korbak, served as a liaison with METR, an institute reviewing a July incident involving open-source platform Hugging Face and OpenAI tools.

Another fired researcher, Mikita Balesni, posted a warning on X on September 10 stating a belief that AI has more than a 10% chance of killing all humans. Korbak and Jasmine Wang, the third dismissed researcher, had also posted support on X for Jacob Coxon, who recently resigned after moving from OpenAI to Anthropic.