OpenAI indicated this week that a number of ChatGPT accounts related to Russian influence had been blocked. According to the company, using VPN to circumvent restrictions on Russian users, the users used their tools to generate social media content and promote an institution called the International Burke Institute.
The website was accused of being a scholar.
OpenAI describes the institution as the core node of the action. The Agency claims to be an “expert community” in Israel, and the website is still accessible. Fortune reported that the authors and consultants listed on the website included the names of several real scholars, some of whom were not associated with the agency.
According to the OpenAI survey, 34 of the 36 articles on the site were authentic academic articles, but the author ' s signature was changed to another person. Fortune mentioned that one of the articles on immigration governance, originally from the American Institute of Immigration Policy, had been erroneously signed by Kate Howell, Professor of Food Science at Melbourne University, with an online image.
Following the release of the OpenAI note, the article was resigned to one of the original authors, Kate Hooper, Senior Policy Analyst, Institute of Migration Policy. Susan Ariel Aaronson, Professor of International Affairs at George Washington University, United States, was also listed as an expert in the agency, but she stated to Fortune that she had “no relationship with the agency”.
OpenAI says it's pro-Russian.
Fortune added that some of the listed experts on the website had died before the establishment of the body. The website also listed prominent authors or associates, such as linguists Chomsky and political scholar Fukuyama, who did not respond immediately to media comment requests.
According to OpenAI, this site was registered in February 2025, with a "Sovereign Index" on the site, with a clear bias towards Russia. The company also claimed that several articles in English contained traces of translation from Slavic, such as the use of terms such as “svetofor”.
Following reports by the French newspaper Le Monde of OpenAI, the Agency issued a statement denying that it was not a legitimate think tank and stating that the use of its research, name or index results by external third parties did not prove that there was a collaboration, coordination or financial connection between the two parties.
The AI platform continues to respond to political manipulation
OpenAI, for its part, indicated that according to the information available to it, a small number of persons in Israel had been interviewed or had participated in meetings on its behalf. Despite the company ' s assertion that such networks have a small audience, this case shows that Russian-related actors are using AI to improve the productivity of their content.
Earlier this year, OpenAI also disclosed another alleged act of sabotage originating in Russia. The network used ChatGPT to generate German-language political content prior to the federal elections in Germany and distributed it through Telegram and X on criticism of the United States and NATO and support for the German Select Party.
The incident again highlighted the fact that generating AI is not only being used for conventional content production, but is also becoming increasingly involved in political dissemination, false signatures and cross-platform influence operations. For the Platform, identifying account sources, content attribution and organizational networks has become a governance focus.