OpenAI said Tuesday it had banned a cluster of ChatGPT accounts originating in Russia that were being used to support a covert influence campaign built around a fake Israeli think tank, describing it as one of the more elaborate operations it has disrupted since Russia’s invasion of Ukraine began.
Building a Fake Institution
The operation centred on the International Burke Institute, a self-described “expert community” whose website listed a street address in Israel and falsely claimed to feature prominent academics including Francis Fukuyama and Noam Chomsky among its contributors. In a review of 36 articles linked to purported experts on the site published between September 2025 and May 2026, OpenAI found that 34 had been copied from elsewhere on the internet, sometimes with incorrect attribution. In one case, a paper on migration policy was falsely attributed to a food sciences professor from Australia.
The institute also published what it called a “sovereignty index,” a metric that rated countries in a way designed to cast Russia favourably while criticising Western nations. Reports targeting France, Germany, the United States and the European Union were, in OpenAI’s assessment, openly polemical rather than analytical.
How the Operation Evaded Detection
OpenAI does not allow ChatGPT access from Russia, so the operators routed their connections through VPNs to disguise their location. Working primarily in Russian, they used the model to draft English-language social media content that was then posted on Substack, Telegram, X, Facebook and LinkedIn, and specifically directed ChatGPT to strip out any linguistic features that might reveal the content had been produced by Russian speakers.
“Although the campaign appears to have reached relatively small audiences, its elaborate construction distinguishes it from other Russia-linked influence operations we have disrupted,” OpenAI said in a blog post detailing the takedown. The company said the significance of the operation lay less in the size of the audience it reached than in the infrastructure it had built, an otherwise credible-looking institution complete with purported experts, republished academic work and a proprietary index.
Part of a Wider Pattern
OpenAI said it had also separately banned Chinese-speaking users who used ChatGPT to generate anti-Trump slogans, cartoons and multilingual social media comments as part of a different influence effort, and noted that it has disrupted similar Russia-linked operations before, including accounts connected to Rybar, a pro-Russia media outlet the UK government has described as being partially coordinated by the Russian presidential administration and funded by the Russian state.
The company said none of the content published directly on the International Burke Institute’s own website showed evidence of having been generated using ChatGPT, with the AI-generated material confined to the social media posts promoting the site.
Why It Matters
The takedown adds to a growing body of evidence that state-linked actors are using generative AI tools to lower the cost of building convincing disinformation infrastructure, from fabricated expert credentials to automated, stylistically laundered social media content. OpenAI has said it will continue to publish findings on influence operations it disrupts as part of an effort to track how state actors are adapting their tactics to AI tools.
