OpenAI Bans Russian ChatGPT Accounts Used in Coordinated Influence Operation

Key Findings

• OpenAI banned Russian ChatGPT accounts operating a fake think tank called the International Burke Institute (IBI) to spread pro-Russia narratives

• Operators used VPNs to bypass Russian access restrictions and prompted ChatGPT in Russian to generate English-language social media content

• The operation centered on a fake sovereignty index that consistently ranked Russia favorably while criticizing Western nations

• The website contained 34 of 36 sampled articles that were copied or misattributed from legitimate academic sources

• The campaign reached relatively small audiences but demonstrated a sophisticated approach to manufacturing institutional credibility

Background

OpenAI discovered the operation while investigating AI-generated social media posts that led to a broader influence campaign. The International Burke Institute presented itself as an Israel-based expert community when it was actually operated from Russia. The website launched in February 2025 and claimed associations with prominent figures like Francis Fukuyama and Noam Chomsky to establish legitimacy.

The Operational Method

The campaign did not rely on dramatic deepfakes or viral bot armies. Instead, it followed a more conventional playbook: create something that looks credible, fill it with legitimate-seeming content, then use social media to funnel people toward it. The operators generated promotional posts and comments across Substack, Telegram, X, Facebook and LinkedIn, all pointing back to the IBI website and its content.

To avoid detection, operators instructed ChatGPT to hide linguistic clues revealing Russian origins while conducting all their prompts in Russian. They used VPNs to access OpenAI's platform, which restricts usage from Russia.

The Fake Think Tank Infrastructure

The IBI website presented itself as a legitimate research institution complete with expert profiles and academic articles. However, investigation revealed most published material was plagiarized. One article about the China-Pakistan Economic Corridor was copied from Cambridge University Press but falsely attributed to a professor at the University of Nottingham. Some pieces appeared to be written by Slavic speakers and then machine-translated into English.

OpenAI noted that ChatGPT was not used to generate the website content itself, but rather to create the promotional layer that drove traffic to it and made the operation appear more organic and active.

The Sovereignty Index

The centerpiece of the operation was the so-called Sovereignty Index, also called the Burke Index. It ranked countries across seven dimensions: political, economic, technological, information, cultural, cognitive and military sovereignty. Each dimension was scored up to 100 points, combining to a maximum score of 700.

The index consistently placed Russia favorably while criticizing Western states, particularly France, Germany, the EU and the United States. Russia ranked fourth overall with 601.4 points, positioned as a stable world power with strong military and economic self-sufficiency. The U.S. topped the list at 650.9 but the index framed Western nations negatively.

Made-up indices are effective propaganda tools because numbers carry inherent authority. When paired with vague methodologies, professional formatting and academic-looking articles, such rankings can travel far beyond what anonymous posts could achieve.

Broader Campaign Elements

Beyond IBI promotion, operators ran a German-language Telegram channel called "Lahme Ente" (Lame Duck) that attacked Ukraine, the EU and the German government while advocating for closer Russian ties. ChatGPT was used to create logos for channels targeting Germany, France, Poland, Turkey and the United States.

One operator repeatedly asked ChatGPT to generate Russian-language summaries of activity from a channel called "American Observer," suggesting plans to monitor and potentially scale operations.

Significance and Implications

OpenAI stated the operation's importance lay less in the audience it reached than in the infrastructure it built. While the organization attracted relatively small audiences with individual Telegram channels drawing 10,000 to 20,000 followers, it demonstrated how AI can be weaponized as a supporting tool within broader influence campaigns.

The case illustrates a concerning trend: influence operators using AI to manufacture authority, obscure narrative sources and create assets that could be expanded over time. It also demonstrates how such operations can be exposed through the digital fingerprints they leave behind.

Sources

• https://securityaffairs.com/197878/intelligence/openai-banned-russian-chatgpt-accounts-backing-covert-influence-operation.html

• https://thehackernews.com/2026/08/openai-bans-russian-chatgpt-accounts.html

• https://www.gate.com/news/detail/openai-bans-russian-chatgpt-accounts-used-in-covert-misinformation-campaign-23710991

• https://www.yahoo.com/news/politics/articles/openai-bans-russian-chatgpt-accounts-121927726.html

• https://finance.yahoo.com/technology/ai/articles/microsoft-backed-openai-bans-russian-121812962.html

• https://www.techmeme.com/260825/p13