Russian pranksters, posing as head of Ukraine's Foreign Intelligence Service Rustem Umerov, held a 17-minute video call with three high-ranking OpenAI officials responsible for state and international security. During the conversation, the parties discussed military support for Ukraine, the provision of weapons for offensive operations, and the use of AI technologies for defense purposes. The recording of the conversation was published on the Rumble platform, and the fact of the incident was confirmed by OpenAI spokesperson Nate Evans, who announced that an internal investigation had been conducted. This was reported by RBC-Ukraine with reference to The Intercept.
Who fell into the trap: three key OpenAI figures
Posing as “Umerov,” the pranksters managed to convince three OpenAI executives, each of whom has significant experience in national security. Sasha Baker leads national security policy at the company and previously held a high-ranking position in the U.S. Department of Defense. Caroline Zir heads the strategic initiatives directorate and also has Pentagon experience. Ben Crampton is responsible for the company's European national security engagement; prior to joining OpenAI, he worked in United Nations and European Union structures. The combined profile of these specialists—former Pentagon official, former UN/EU diplomat, and national security director—makes the incident particularly telling in terms of the vulnerability of corporate communication channels.
What was discussed: weapons, Daybreak Blue, and the “Washington colleague”
According to the published recording, the parties discussed the possibility of military aid to Ukraine, including the provision of weapons for offensive operations and AI technologies for defense needs. Crampton noted that the company strives for a “deep relationship with Ukraine,” yet within the organization, there is caution regarding involving technologies in offensive military purposes. Separately, joint initiatives with the Ukrainian government were discussed: in particular, using the Daybreak Blue model to find and fix vulnerabilities in computer systems, as well as expanding cooperation with the Ukrainian Ministry of Defense through a “Washington colleague” who previously worked with the Delta unit of the Ukrainian defense department.
Position on Musk, Trump, and Baker's “Regret”
The conversation also touched upon the situation around Elon Musk and the U.S. President's stance on Ukraine. Sasha Baker assured her interlocutor that the majority of Americans and company researchers support Ukraine. Answering questions about Donald Trump's statements toward Volodymyr Zelenskyy, Baker noted that such rhetoric is “regrettable.” These remarks, made during a conversation that turned out to be conducted not by Umerov, but by a Russian prankster, became one of the most cited fragments of the published recording.
OpenAI's reaction and corporate security questions
OpenAI spokesperson Nate Evans confirmed the incident and stated that the company had conducted an internal investigation involving competent authorities. According to him, measures have been taken to prevent such situations in the future. The company did not disclose specific details of the investigation, nor exactly what data was transmitted to “Umerov” during the 17-minute conversation. The incident once again raises the question of how vulnerable corporate video conferences and communication channels of tech companies are to targeted social engineering, especially when national security and defense cooperation are at stake.
