Russian Pranksters Fool OpenAI Security Team in Fake Call

Three of OpenAI’s most senior national security staff sat through a 17-minute video call about arming Ukraine. The man on screen was not who they thought he was.

A Russian prank duo had posed as Ukraine’s foreign intelligence chief, and the recording is now public. Here is what happened, what OpenAI says, and why this matters far beyond one embarrassing call.

What Exactly Happened?

The story was first reported by The Intercept on 11 November 2026. Here are the confirmed and reported details.

  • Who pulled the prank: The Russian duo known as Vovan and Lexus (Vladimir Kuznetsov and Alexei Stolyarov). They have run political prank calls since 2011 and have been accused of working with the Russian government.
  • Who they posed as: Rustem Umerov, head of Ukraine’s Foreign Intelligence Service.
  • Who was fooled: Sasha Baker (head of national security policy), Caroline Zier (lead for strategic initiatives) and Ben Crampton (who leads European national security work) at OpenAI.
  • Length: About 17 minutes. The duo posted the recording online.
  • How it got arranged: The pranksters say they first contacted a senior Ukrainian official in San Francisco while posing as Umerov. That official reportedly believed the request was real and offered to set up a meeting with OpenAI.

The full report is on The Intercept.

What was discussed on the call

According to reporting based on the recording, the OpenAI staff talked about possible support for Ukraine’s defense, including whether OpenAI could offer offensive military AI uses. They also discussed the company’s Daybreak Blue security model and early contacts with Ukraine’s Defense Ministry. Baker reportedly said OpenAI had been in touch with the Pentagon about helping Ukraine.

It is important to be precise here. The call did not prove that OpenAI has supplied any offensive AI tools to Ukraine. It shows what staff were willing to discuss with someone they believed was an official.

What OpenAI says

OpenAI spokesperson Nate Evans confirmed the incident to The Intercept. He said the company investigated with relevant authorities weeks ago, concluded the call was a prank, and took steps to prevent similar situations. No compromise of data or systems has been reported.

Why It Matters

For Users

Nothing suggests your ChatGPT data was touched. But the story is a reminder that AI companies now work on defense and national security topics, and the people running those talks can be targeted like anyone else.

For Businesses

This is a textbook social engineering case. The attackers did not hack anything. They borrowed trust from a real person, who vouched for a fake one. Any company that takes high-level meetings from outside parties can fall into the same trap.

The lesson is simple: verify the caller through a second channel before you share anything, even when the introduction comes from someone you trust.

For the AI Industry

AI labs are moving closer to governments and militaries. That raises the value of their staff as targets for intelligence gathering and influence operations. If pranksters can get 17 minutes with a security team, a serious adversary might get more.

The incident also adds to the pressure on AI companies to be open about their government work. We covered the political side of that in our post on AI CEOs possibly testifying under oath.

Social Media Reaction

We could not verify individual posts well enough to quote them, so this section summarizes the public debate in news coverage and commentary.

Critics say the story is a serious security failure. Futurism argued that if a pair of pranksters can reach senior staff, a malicious actor could do the same. Others point out that national security staff should know to verify identity before talking about defense cooperation.

Defenders note that OpenAI says it investigated weeks before the story came out and that no data compromise has been reported. Some also point out that Vovan and Lexus have fooled politicians and officials in several countries, including a UK defense secretary in 2022, so OpenAI is not the first.

Skeptics of the pranksters remind readers that this duo has been accused of working with Russian state interests. A call posted online is also edited for impact, so treat the recording as one side of the story.

What Should You Do Next?

  1. Add a second-channel check. Before any sensitive meeting, confirm the other person through a known phone number or official email, not the link you were sent.
  2. Train for introductions. Teach your team that a warm intro is not proof of identity. Ask who vouched for whom, and verify that chain.
  3. Limit what you say on first calls. Treat first contacts as listening calls. Do not discuss partnerships, contracts or internal limits until identity is confirmed.

FAQ

Who are Vovan and Lexus?

They are a Russian prank duo, Vladimir Kuznetsov and Alexei Stolyarov, known for calling officials while impersonating other politicians. They have done this since 2011.

Did OpenAI give military AI to Ukraine?

The call does not show that. Staff discussed possible support and contacts, but no supply of offensive AI tools has been confirmed.

Was OpenAI hacked?

No hack has been reported. This was social engineering, meaning people were tricked, not systems.

How did the fake call get set up?

The pranksters say a senior Ukrainian official in San Francisco believed they were Umerov and offered to arrange the OpenAI meeting.

Conclusion

The OpenAI prank call is less about comedy and more about trust. As AI companies enter defense work, verifying who is on the other end of the line becomes a core security skill. Related policy news: see our earlier report on how the US is blocking China from renting Nvidia GPUs abroad.

For daily, plain-English AI news, follow AI Trends Setter.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest articles

Related articles