Fired OpenAI Researchers Warn of Chilling Internal Culture

The trio, who worked on AI alignment and safety, addressed the letter to OpenAI’s safety committees, arguing that the company cannot ensure AI security in isolation. The researchers contend that their firings, which were delivered with only verbal explanations, contradict the company’s stated commitment to transparency. Balesni noted he was informed the company no longer trusted him due to his interactions with external safety bodies, specifically the nonprofit METR. Korbak emphasized that communicating with such organizations was a core requirement of his role, not a violation of policy.

OpenAI maintains that the departures followed an internal investigation into the handling of sensitive information. A company spokesperson stated the terminations were unrelated to the employees' efforts to raise safety concerns. Despite this, the former staff members fear their exit serves as a pretext to restrict collaboration with external auditors. In their letter, they urged OpenAI to formally embed third-party oversight and maintain an open dialogue with the broader safety ecosystem to prevent further erosion of the company’s collaborative standards.