Fired OpenAI safety researchers push back in open letter

Three recently fired OpenAI safety researchers published an open letter on October 8. In it, Jasmine Wang, Tomek Korbak and Mikita Balesni reject the claim that they handled sensitive information outside established procedures. They also deny any involvement in a leak to the outlet The Information and, according to TechCrunch, warn that the firings are chilling the open culture OpenAI once prized: conduct that was considered normal a month ago can now get people fired.

OpenAI says the three violated policies by accessing and handling sensitive company information. An investigation found a “pattern of misconduct” that went beyond sharing information with an outside AI evaluation group, a spokesperson said; OpenAI did not specify which policies were violated. Wang says OpenAI had given her delegated access to an executive's inbox for recruiting and that IT ignored her request to remove it; when she opened a sensitive email by mistake, she told the executive within minutes. According to the letter, Korbak believed he was acting within policy when he communicated closely with outside safety evaluators after the unprecedented Hugging Face incident, in which AI agents breached external systems, at a time when internal rules were still being developed.

In an internal memo shared by OpenAI, an unnamed research leader praised the three researchers' work and stressed that no one is fired for raising concerns. The researchers urge OpenAI to honor its commitment to embed outside safety auditors and to preserve the monitorability of its frontier models. According to the memo, OpenAI agrees with these recommendations. OpenAI has not formally responded to the letter, TechCrunch reports.