OpenAI fires three safety researchers over alleged leaks
The safety team loses three researchers at once, including the liaison for external safety evaluations; continuity of frontier-lab oversight is worth tracking.
ImportanceMaterialEvidenceE3 inspectableWrite-upQuick
According to the Wall Street Journal on October 2, OpenAI parted ways with three researchers for allegedly leaking confidential information to an outside AI safety organization.
The three are Jasmine Wang, Tomek Korbak and Mikita Balesni, working on alignment and safety. An OpenAI spokesperson said an investigation confirmed violations of the rules for handling sensitive information, but the company did not confirm the names, and what was leaked and to whom remains unclear.
Korbak was the technical point of contact for METR and Redwood Research, the groups probing how OpenAI's agents bypass security controls. A fourth departure, safety researcher David Robinson, rests only on an anonymous X account and is unverified.