OpenAI Uncovers Russian Influence Campaign Promoting Fake Israeli Research Institute

OpenAI revealed a Russian influence campaign that exploited its AI capabilities to promote a fictitious Israeli research institute called the International Burke Institute (IBI). The campaign was uncovered after OpenAI detected and blocked several ChatGPT accounts operated by individuals writing prompts in Russian to generate posts and comments across platforms like Substack, Telegram, X (formerly Twitter), Facebook, and LinkedIn. Most content was in English, with efforts made to conceal the Russian origin of the activity. OpenAI stated it does not permit access to its models from Russia and noted the use of VPNs to bypass restrictions.

The IBI website claimed to be based in Israel and featured numerous academic articles, many of which were plagiarized from genuine publications with incorrect attributions. Some articles were originally written in Slavic languages and automatically translated into English. The site’s most prominent feature was a "Sovereignty Index" that praised Russia and criticized Western countries. ChatGPT was primarily used to generate posts promoting the fake institute, some published under accounts bearing the IBI name and logo, while others appeared to be from fake regular users mainly posting IBI content. Operators also generated comments on real users’ posts, often urging followers to subscribe to IBI channels.

One operator posted in German on the Telegram channel Lahme Ente (Lame Duck), criticizing Ukraine, the European Union, and the German government. Another created logos for about a dozen Telegram channels focused on Germany, the US, France, Poland, and Turkey, some of which published IBI content. OpenAI described the AI activity as supplementary to the core campaign of running the fake institute, which falsely claimed to be an independent expert community based in Israel with renowned members like Francis Fukuyama and Noam Chomsky. A sample review of 36 articles found 34 were copied from other sources, some years old and misattributed, indicating an attempt to lend legitimacy to IBI through authentic content.

The IBI site also included reports ranking countries’ sovereignty, mostly critical of Western nations. For example, the US was criticized for actions under Trump perceived as harming autonomy, and France was described as "opening like a safe and being sold off in parts" after a decade of Macron’s rule. OpenAI found online evidence that some Israeli individuals mentioned or represented IBI in articles and conferences but could not clarify the relationship between these parties, IBI, and the Russian operators. Despite the campaign’s limited reach, IBI posts had low views and few subscribers, with Telegram channels having 10,000 to 20,000 followers, OpenAI emphasized the campaign’s significance lies in the infrastructure it built. The use of AI-generated posts helped create a seemingly credible institution with expert backing and unique metrics, illustrating how influence actors can leverage AI to establish authority, obscure origins, and build assets for long-term narrative expansion.