OpenAI Uncovers Russian Operation Masquerading as Israeli Think Tank
The campaign used ChatGPT to generate social media content, promote pro-Russia narratives and obscure the operators’ origins
Photo by Koichi Mitsui/AFLO via Reuters Connect
OpenAI says it has disrupted a previously unreported Russian influence operation that used ChatGPT to promote what appeared to be an Israeli think tank across social media platforms.
In a lengthy article on its website published yesterday (Tuesday), the company said it had banned a cluster of ChatGPT accounts that it believes originated in Russia and were used to generate posts for the International Burke Institute (IBI). The organization presented itself as an expert community based in Israel, but OpenAI's investigation found evidence pointing to Russian operators behind the campaign.
The operation used ChatGPT to produce English-language posts and comments promoting IBI content on X, LinkedIn, Facebook, Substack and Telegram. The operators also instructed the model to remove linguistic clues that could reveal they were Russian. Because OpenAI does not provide access to its models from Russia, the accounts used VPNs to connect to ChatGPT.
The IBI website, registered in February 2025, presented itself as a research organization and claimed affiliations with prominent academics. OpenAI examined 36 articles attributed to experts on the site and found that 34 had been copied from elsewhere, in some cases with incorrect authorship.
The site also promoted a so-called “sovereignty index” that portrayed Russia favorably while ranking Western countries negatively.
The operation extended beyond IBI. One operator used ChatGPT to generate German-language content criticizing Ukraine, the European Union and the German government while calling for closer relations with Russia. Another generated logos and other material for Telegram channels posing as media outlets focused on countries including the US, Germany, France, Poland and Türkiye.
OpenAI said the campaign appeared to have reached relatively small audiences, but its structure was more elaborate than previous Russia-linked influence operations it has disrupted.
The case illustrates how generative AI can be used not necessarily to create sophisticated propaganda, but to scale the routine work of a covert influence campaign — producing posts, comments, translations and branding while helping operators maintain the appearance of independent voices.