Three senior OpenAI executives responsible for the national security and defense direction became victims of a prank by the Russian prankster duo "Vovan and Lexus" (Volodymyr Kuznetsov and Oleksiy Stolyarov). During a 17-minute video call, in which the Russians posed as a Ukrainian official, company representatives openly discussed the prospects of military assistance to Ukraine and internal discussions regarding the use of AI in offensive operations.
Key OpenAI security specialists with significant experience in government structures contacted the fake Ukrainian official: Head of National Security Policy Sasha Baker, Head of Strategic Initiatives Caroline Zier, and European National Security Lead Ben Crampton. Before joining the company, Baker and Zier held high-ranking positions at the Pentagon, while Crampton worked in EU and UN structures, reports the American publication The Intercept, citing a recording of the conversation released by pranksters.
According to one of the pranksters, organizing the meeting turned out to be easy. The Russians first contacted a Ukrainian official in San Francisco on behalf of Rustem Umerov. Since this official was just establishing contacts with OpenAI through the California diaspora and had known Sasha Baker since her time at the Pentagon, he helped organize a video conference. During the call itself, the interlocutor posing as Umerov kept his camera off.
What was discussed
During the conversation, OpenAI representatives shared details of both the current cooperation with Ukraine and the company's internal restrictions.
Top managers spoke about interaction with the Ukrainian government on the basis of the Daybreak Blue AI model (a system for detecting and eliminating computer vulnerabilities), as well as about the initial partnership with the Ministry of Defense of Ukraine. In particular, contact is maintained through an OpenAI employee in Washington who previously worked with the team of the Delta combat management system.
Responding to a question about the possibility of providing Ukraine with offensive military capabilities based on AI, Ben Crampton stated that the company seeks the deepest possible relations with Ukraine, but it is impossible to implement this instantly. According to him, many OpenAI top managers still do not wish to become too deeply involved in the offensive application of technologies. At the same time, he noted: "Part of our work is to explain that under certain circumstances, offense and defense cannot be clearly distinguished."
Sasha Baker noted that the company is in contact with the Pentagon regarding assistance to Ukraine. Commenting on the cool rhetoric of US President Donald Trump toward Volodymyr Zelensky, she assured the interlocutor that the majority of the American government, US citizens, and OpenAI researchers resolutely support Ukraine. When asked to help counter Elon Musk's influence, Baker responded with a joke that the company would hardly be able to help with that, since Musk "is also not a fan of our CEO, so we are in the same boat."
OpenAI's reaction
In a comment for The Intercept, OpenAI spokesperson Nate Evans confirmed the incident, noting that the company conducted an internal investigation together with the relevant authorities several weeks ago.
“We came to the conclusion that this was a prank organized by individuals who have a history of attacks on officials and other companies. We have taken measures to prevent similar situations from recurring in the future,” a company representative said.
The Embassy of Ukraine in the USA did not promptly respond to the publication's request for comment.