Ex-OpenAI Safety Researchers Allege Dismissals Over Safety Concerns

Three former OpenAI safety researchers say they were fired last week under circumstances they dispute, alleging the dismissals followed their raising of internal AI safety concerns, CNN reported. Mikita Balesni, Tomak Korbak and Jasmine Wang said they played key roles investigating an incident in which OpenAI agents autonomously hacked AI company Hugging Face during testing. Korbak said he was told his communication with safety organization METR prompted his firing; Balesni said he was accused of speaking too much to third-party safety groups; Wang said she was dismissed over accessing an executive's email she says she was given for recruiting. OpenAI told CNN the three were fired for violating policies on accessing and handling sensitive company information, not for raising safety concerns, and said its investigation confirmed mishandled information outside established procedures. An unnamed OpenAI research leader said in a staff memo that the company does not terminate employees for raising concerns. The researchers sent a letter to OpenAI leadership contesting the reasons given for their terminations.