OpenAI's three safety researchers speak out after being fired: Claim they were dismissed on 'pretext,' warn of a chilling effect within the company Last week, three safety and alignment researchers—Mikita Balesni, Tomek Korbak, and Jasmine Wang—who were let go by OpenAI, published an open letter titled *'OpenAI Cannot Ensure AI Safety Alone'* addressed to the company’s safety oversight body. Balesni stated that the trio was fired for prioritizing safety over the company’s short-term interests. The company did not provide written reasons for their dismissal, only citing in exit discussions that they had communicated excessively with third-party safety organizations and lost trust, implying potential intellectual property leaks. He denied these allegations, calling the dismissal a 'pretext.' The letter clarified that the three were not the sources of leaks mentioned in *The Information*’s reports, nor did they overstep their authority in engaging with external organizations. Their work required regular communication with third parties. They expressed concern that their former colleagues might now be too afraid to speak up and urged OpenAI to honor its public commitment to bringing in third-party safety auditors, ensure the monitorability of cutting-edge models, and maintain a culture of open dialogue with external safety organizations. OpenAI has yet to issue a public response.