Former OpenAI Safety Researchers Deny Misconduct Claims and Warn of Cultural Shift
Three former safety researchers at OpenAI—Jasmine Wang, Tomek Korbak, and Mikita Balesni—have published an open letter denying allegations that they mishandled sensitive information. The trio was fired last week for allegedly sharing confidential data with a third-party AI safety organization. In their statement, the researchers claim their dismissal is part of a broader cultural shift at OpenAI, where previously encouraged collaboration and transparency are now being replaced by fear and uncertainty. The letter warns that this change in company culture could have long-term negative effects on internal communication and external accountability efforts related to AI safety. The researchers also denied involvement in an alleged leak about the limitations of monitoring new models' reasoning processes. OpenAI has not directly responded but shared a memo from a research leader stating the firings were unrelated to raising concerns or speaking out. The incident comes amid growing scrutiny over recent safety issues at OpenAI, including leaks and technical breaches involving AI agents.
Topics: AI Safety Corporate Culture Open Communication Workplace Ethics