OpenAI Responds to Safety Researcher Resignation: Due to Sensitive Information Violation, Not Suppression of Safety Concerns

On October 9, a head of research at OpenAI stated that the company has ended its collaboration with three security researchers—Jasmine, Mikita, and Tomek—following an internal investigation into violations of policies regarding the handling of sensitive information. OpenAI noted that the issues uncovered during the investigation exceeded those disclosed in the researchers' open letter, leading the company to uphold its decision to terminate the employment relationship.

The company emphasized that this action was not motivated by the researchers raising AI safety concerns or expressing differing opinions, clarifying that internal safety discussions and research debates have consistently been part of daily operations. Regarding the external oversight mentioned in the open letter, OpenAI stated it is advancing cooperation with third-party security assessment bodies and will publish details in the coming weeks. The company agrees that ensuring the monitorability of frontier models requires a collective commitment across the industry, with monitorability long serving as a core component of its research initiatives. (Jin10)