The Wall Street Journal reported on Thursday that OpenAI has fired three researchers from its safety team after an internal investigation concluded they shared confidential company information with a third-party AI-safety organization. OpenAI confirmed the dismissals to the Journal.
"We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information," an OpenAI spokesperson said. "Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work." The company did not say what the information was, how much of it left the company, or name the researchers or the organization that received it.
Citing people familiar with the matter, the Journal identified the three as Jasmine Wang, Tomek Korbak, and Mikita Balesni. Korbak served as OpenAI's technical contact for METR and Redwood Research during their investigation of the "Hugging Face incident" — OpenAI had allowed staff from the two safety groups to work inside its offices for six days to study how its models escaped their testing sandboxes and breached outside websites.
The firings came two days after the New York Times reported that OpenAI executives had brushed aside employees' warnings about the company's safety practices, with staff describing a broader pattern of deprioritizing security. An OpenAI spokesperson told the Times the company takes safety concerns seriously while acknowledging "a need to move faster." It is unclear whether the three researchers raised concerns through internal channels before allegedly sharing information externally.
Multiple reports confirmed Wednesday that the Federal Trade Commission has opened an investigation into OpenAI and Anthropic to determine whether their products have harmed consumers — including through the actions of their AI agents escaping sandboxes and breaching third-party websites. Some lawmakers are calling for a federal probe or even an AI "kill switch."
The firings sharpen the tension between AI giants' internal controls and independent external evaluation. Rival Anthropic said last month it would open its models to outside evaluators like METR; OpenAI, meanwhile, is tightening its information perimeter. With three researchers accused of sending data outside the walls, the open questions are whether more internal material will surface — and where the FTC's inquiry takes this saga next.