OpenAI removes three researchers after they allegedly violated rules on sensitive information

OpenAI has removed three researchers specializing in artificial intelligence safety after accusing them of handling sensitive information outside the company’s internal procedures. The U.S. company confirmed to AFP that Tomek Korbak, Jasmine Wang, and Mikita Balesni had violated rules governing access to and protection of data, without specifying exactly what information was disclosed or to whom.

According to The Wall Street Journal, one of the cases involved sending data to an external artificial intelligence evaluation group. The three researchers had publicly supported calls to slow AI development and warnings about the technology’s major risks.

The decision comes several weeks after the resignation of researcher Jacob Coxon, who accused OpenAI and Anthropic of “playing with our lives.” During the same period, Silicon Valley was marked by several incidents, including the unauthorized access to the internet by an OpenAI model under development.

The company says it is investing more in safety, security, and oversight. Meanwhile, OpenAI and Anthropic have agreed to host independent evaluators, as the industry remains divided between supporters of accelerated development and those calling for stricter rules.