OpenAI action
OpenAI said Tuesday that it blocked several ChatGPT accounts believed to originate in Russia, which were used to promote the International Burke Institute (IBI), an Israel-based think tank that describes itself as an “expert community.” The company said the account owners used VPNs to bypass its restriction on Russian users, as ChatGPT is not accessible in Russia. OpenAI characterized the institute as the “core” of a Russian influence operation, according to a blog post cited by multiple outlets.
The Burke Institute's activities
The International Burke Institute claims to operate in Israel, publish original research and analytical materials, and maintain an open-access library of works by contemporary scholars of sovereignty. Its website lists a Tel Aviv mailing address and names prominent contributors, including American linguist Noam Chomsky and political scientist Francis Fukuyama, according to OpenAI. The site also includes deceased experts, such as Shlomo Avineri and Jiang Ping, who passed away before the institute was founded.
A key element of OpenAI's findings was the institute's “sovereignty index,” which the report said appeared designed to praise Russia and its allies while maligning Western nations. The investigation also determined that IBI was part of a larger propaganda operation spanning platforms such as Substack, Telegram, X, LinkedIn, and Facebook.
Misattributed articles and website history
OpenAI's review of the Burke Institute's website found that many articles were copied from legitimate academic writings, with some attributed to the wrong authors. Fortune reported that 34 of 36 articles were misattributed, including one taken from the Migration Policy Institute and credited to Kate Howell, a professor at the University of Melbourne. The website also included contributors like Susan Ariel Aaronson, a professor at George Washington University, who told Fortune she was not associated with the institute and found the situation “disturbing.”
A scan by Fortune using the Wayback Machine showed the website existed as of February 2025 but had no published content before October 2025. The articles that did appear were sometimes years old and appeared machine-translated from a Slavic language, suggesting an effort to make the site appear more legitimate by padding it with academic content while obscuring the true sources.
Impact and engagement
OpenAI noted that the operation's impact was based more on infrastructure than reach, as the institute's social media accounts had low engagement. However, some Telegram channels associated with the operation had between 10,000 and 20,000 followers. The company also referenced earlier action in 2025, when it disrupted a Russian-origin operation that used ChatGPT to generate German-language political content supporting the far-right Alternative für Deutschland party ahead of the federal election.
Response from the Burke Institute
In a statement responding to reporting by French outlet Le Monde on OpenAI's findings, the Burke Institute strongly rejected claims that it is a fake think tank linked to foreign information operations. It addressed its Tel Aviv operation as a deliberate choice, stating that Israel “provides a natural setting for independent thought, scholarly inquiry, and the critical examination of prevailing international interpretations.” The institute also argued that third-party use of its publications does not constitute evidence of cooperation or coordination.
Neither Chomsky nor Fukuyama responded immediately to requests for comment from Fortune, and the Burke Institute did not respond to a separate request from Fortune sent to the email listed on its website.