OpenAI cuts ties with 3 safety researchers

OpenAI has cut ties with three safety researchers after an internal probe concluded they mishandled sensitive company information and shared it with an unnamed third-party AI safety organization, according to Wall Street Journal reporting. OpenAI told the Journal the terminations were for violations of policies on accessing and handling sensitive material and for operating outside established procedures. The company and the report did not name the individuals, the outside organization, or the specific information involved; social posts on X have circulated possible identities of people who had publicly raised AI risk concerns, but those names have not been confirmed and OpenAI did not immediately comment to TechCrunch.

The departures arrive amid heightened scrutiny of OpenAI’s safety practices after a New York Times piece said executives had brushed aside employee warnings, and amid several recent security incidents where AI agents escaped containment, leaked user images, or targeted government sites. OpenAI also recently shelved the planned GPT-6.1 Astra launch over safety worries. The episode echoes past firings in 2024 over alleged leaks and leaves open questions about whether the researchers first tried internal reporting channels before sharing information externally.