OpenAI has banned a cluster of Russian ChatGPT accounts behind a covert influence operation promoting a fake Israeli think tank called the International Burke Institute. Operating via VPNs to bypass regional blocks, the actors used Russian prompts to generate English social media posts across platforms like X and Telegram. While the campaign had limited reach, it utilized plagiarized academic papers and a biased 'sovereignty index' to manufacture authority and support pro-Kremlin narratives.
ChatGPT account bans
- ▪The banned ChatGPT accounts were operated by individuals in Russia who used virtual private networks to bypass OpenAI's regional access restrictions.
- ▪OpenAI banned a cluster of Russian ChatGPT accounts on August 25, 2026, that were linked to a covert online influence campaign.
International Burke Institute operation
- ▪The Russian influence campaign promoted the International Burke Institute, a fictional organization presenting itself as an Israel-based think tank.
- ▪The International Burke Institute website featured a 'sovereignty index' that scored Russia favorably while criticizing Western nations that opposed its invasion of Ukraine.
- ▪Out of 36 articles sampled from the International Burke Institute website, 34 were copied from legitimate academic sources, including some misattributed to Noam Chomsky and Francis Fukuyama.
Russian influence campaign tactics
- ▪The operators used Russian-language prompts to generate English social media posts and comments on platforms including Substack, Telegram, X, Facebook, and LinkedIn.
- ▪The operators instructed ChatGPT to remove linguistic clues that would reveal they were Russian speakers, though some posts still contained awkward literal translations like 'Svetofor coalition' for Germany's traffic-light coalition.
Limited audience reach
- ▪The immediate impact of the International Burke Institute campaign was limited, with individual social media posts receiving low view counts and official accounts maintaining low subscriber numbers.
- ▪OpenAI categorized the International Burke Institute campaign at the lower end of Category Three on the Brookings Breakout Scale, noting its Telegram channels had between 10,000 and 20,000 followers.
OpenAI threat disruption efforts
- ▪OpenAI previously disrupted other Russia-linked operations, including the 'Doppelganger' campaign in 2024 and accounts connected to the state-funded media company Rybar in February 2026.
- ▪OpenAI's Intelligence & Investigations team has disrupted more than 40 malicious networks since the company began publishing public threat reports in early 2024.
Story comments
Loading comments…