OpenAI fired three researchers from its safety and alignment teams, Jasmine Wang, Tomek Korbak, and Mikita Balesni, in part for allegedly sharing confidential material with an outside AI safety group, according to the Wall Street Journal. OpenAI said its own investigation found the three had broken its rules on handling sensitive information. Korbak had previously described himself as OpenAI’s technical contact for METR and Redwood Research, the two outside safety groups OpenAI brought in for six days of on-site work investigating the Hugging Face breach. The firings come as AI labs face pressure to open their models to independent evaluators, and Anthropic agreed last month to let METR check its safety practices.
