OpenAI has taken down a network of Russian accounts on ChatGPT that were used to carry out a disinformation campaign and manipulate public opinion. The attackers used artificial intelligence to generate content on social media, promoting a fake “think tank” supposedly from Israel.
OpenAI has taken down a network of Russian accounts on ChatGPT that were used to carry out a disinformation campaign and manipulate public opinion. The attackers used artificial intelligence to generate content on social media, promoting a fake “think tank” supposedly from Israel.
OpenAI reported on the blocking of the Russian influence operation in its official report.
According to the company, Russian operators used VPNs to access the platform. They made requests in Russian but instructed the neural network to generate posts in English and German, specifically instructing the AI to hide any linguistic clues that the authors were Russian. The generated posts were published on X, LinkedIn, Facebook, Substack, and Telegram.
The bot farm's main goal was to promote the International Burke Institute (IBI) website, a resource registered in early 2025 that posed as an Israeli think tank. To give the site credibility, the attackers massively stole real scientific articles from the Internet (including from Cambridge University Press), changing the names of the authors.
In addition, the site published a so-called "sovereignty index", which praised Russia and harshly criticized Western countries such as the United States, Germany, and France. Some of the texts showed clear signs of poor-quality machine translation from Slavic languages.
"Although the campaign appears to have reached a relatively small audience, its complex structure distinguishes it from other Russia-related influence operations that we have blocked since the beginning of the war in Ukraine," the company emphasized.
After discovering the influence network, OpenAI blocked the Russian accounts.
