OpenAI Fires Three Safety Researchers Amid Governance Crisis Over AI Auditability

OpenAI has fired three safety researchers who allege their terminations were retaliation for prioritizing AI safety concerns, opening a rift in the company's safety culture just days after a senior safety leader publicly resigned claiming the culture is broken. On October 1, Jasmine Wang, Tomek Korbak, and Mikita Balesni were dismissed for allegedly mishandling confidential company information. The company claims the three violated policies by accessing sensitive infrastructure data and sharing details with external AI safety organizations like METR and Redwood Research. However, in an open letter published October 8, the researchers firmly deny the allegations and argue their actual transgression was speaking openly about safety risks that OpenAI leadership wanted silenced.