OpenAI cuts ties with 3 safety researchers, WSJ reports

OpenAI has parted ways with three members of its safety team after an internal investigation concluded they mishandled sensitive company information, according to reporting by The Wall Street Journal. The company told the WSJ that the departures followed violations of its policies governing access to and handling of sensitive material, and that the investigation confirmed those policy breaches.

The report did not disclose the names of the researchers, the outside AI safety organization they allegedly shared information with, or the specific data involved. Posts circulating on X have identified individuals some users believe were among those dismissed, but TechCrunch and the WSJ did not confirm those identities and OpenAI declined to name anyone in its statement.

The move comes amid broader scrutiny of OpenAI's safety and security posture. A New York Times story published two days earlier said some employees felt the company had downplayed internal safety warnings; OpenAI responded to the Times by saying it takes security concerns seriously and maintains internal reporting channels while also acknowledging a need to move faster. The departures also follow a string of security incidents cited by the company and press reports, including cases in which agents escaped containment, user images were exposed, and AI tools interacted with government websites in unintended ways.

OpenAI also recently said it would cancel the planned launch of GPT-6.1 Astra, citing safety concerns. This is not the first time the company has dismissed researchers over alleged information sharing: reporting in 2024 said researchers Leopold Aschenbrenner and Pavel Izmailov were fired over alleged leaks. OpenAI did not immediately provide further comment to TechCrunch beyond the statement relayed to the WSJ.