OpenAI announced on August 25, 2026 that it banned a cluster of ChatGPT accounts that it says very likely originated in Russia and were used in a covert influence operation. According to OpenAI, operators prompted the model in Russian to generate social media comments—mostly in English—posted on Substack, Telegram, X, Facebook and LinkedIn, and instructed ChatGPT to hide linguistic clues indicating they were Russian. Because OpenAI does not allow access from Russia, the operators used VPNs.
Much of the content promoted a self-described 'expert community' called the International Burke Institute (IBI), tied to a website registered in February 2025 that claimed to be based in Israel. OpenAI says the site's articles were not generated by its models; many were copied from real academic writings, sometimes with false attribution, and some appeared machine-translated from a Slavic language. As one example, the report cites the phrase 'the Svetofor coalition' used to describe Germany's traffic-light coalition.
OpenAI describes a 'sovereignty index' promoting Russia and denigrating Western countries as the operation's most notable feature. Operators also generated German-language posts for a Telegram channel called 'Lahme Ente' and logos for around a dozen channels. OpenAI says the campaign appears to have reached relatively small audiences.
- OpenAI banned ChatGPT accounts it says very likely originated in Russia, using VPNs to bypass its Russia access block
- Content promoted the International Burke Institute, a website registered February 2025 claiming an Israel base
- Posts appeared on Substack, Telegram, X, Facebook and LinkedIn; OpenAI says the campaign reached relatively small audiences
What it means for you
This is a report from OpenAI about catching people misusing its tools for propaganda—it does not change how you use AI or what your business should do. It is a useful reminder that AI makes it cheap to mass-produce believable-looking content in any language, including fake 'expert' websites and social media comments designed to look like they came from real people.
Try this
Next time you see an unfamiliar 'institute' or 'think tank' cited online, spend two minutes checking when its website was registered and whether its articles appear elsewhere with different attribution before you trust or share it.
Who should care
Anyone who reads or shares news and commentary online, and small businesses that rely on social proof or third-party 'expert' sources in their marketing or research.
Skip this if
You already treat anonymous online sources with skepticism and verify claims before acting on them—there is nothing here you need to do differently.
Sources: OpenAI — read the original