OpenAI Fires Three Safety Researchers; Employees Claim Retaliation

OpenAI has confirmed the termination of three prominent safety researchers: Tomek Korbak, Mikita Balesni, and Jasmine Wang. A company spokesperson stated that the dismissals were due to violations of policies regarding the access and handling of sensitive company information.

The fired employees published an open letter asserting that the terminations were pretextual and linked to their recent advocacy for stronger model monitorability and external safety audits. Korbak and Balesni were lead authors of the influential Chain of Thought Monitorability paper, while Wang was a signatory of the 'Pacing the Frontier' petition.

OpenAI declined to specify the nature of the alleged breach beyond the initial statement, maintaining that no one was fired for raising safety concerns. The incident has sparked significant debate regarding internal safety culture and employee dissent at leading AI labs.