OpenAI Bans Russian ChatGPT Accounts Behind Fake Think Tank Pushing Pro-Kremlin Narratives

OpenAI Disrupts Russian Disinformation Network Using ChatGPT

OpenAI banned a cluster of Russian ChatGPT accounts on August 25, 2026, after uncovering a covert online influence campaign designed to spread misinformation and pro-Kremlin narratives across multiple social media platforms1

1

3

3

. The operators used VPNs to bypass OpenAI's restrictions on Russian access and generated AI-powered social media posts promoting the International Burke Institute, a fabricated Israeli think tank4

4

. The campaign distributed content across Substack, Telegram, X, Facebook, and LinkedIn, with most posts written in English despite operators prompting ChatGPT in Russian2

2

.

Fake Israeli Think Tank Built on Plagiarized Academic Papers

The International Burke Institute website, registered in February 2025, presented itself as an expert community of leading academics in economics, politics, sociology, and international relations3

3

. OpenAI's investigation revealed that of 36 articles sampled from the site, 34 were plagiarized academic papers copied from real authors1

1

. Some works were falsely attributed to prominent figures including Francis Fukuyama and Noam Chomsky1

1

. In one glaring example, a migration policy paper was credited to an Australian professor of food science2

2

. The site even listed experts who had passed away before the institute's founding, such as Shlomo Avineri and Jiang Ping5

5

.

Sovereignty Index Designed to Manufacture Authority

The fake Israeli think tank promoted a "Burke Sovereignty Index" that ranked countries based on seven indicators spanning political, economic, technological, informational, cultural, cognitive, and military sovereignty3

3

. The index consistently portrayed Russia favorably while criticizing Western countries supporting Ukraine1

1

. Russia scored 601.4 points, positioning it above many Western nations3

3

. The index used language "typically critical, veering into the polemic" against countries opposing Russia's war in Ukraine2

2

. This sovereignty index served as a tool to manufacture authority and obscure sources of pro-Kremlin narratives4

4

.

AI-Generated Social Media Posts Revealed Russian Origins

The Russian operators instructed ChatGPT to hide linguistic clues indicating their origins, yet obvious tells remained in the AI-generated social media posts1

1

. One article referred to Germany's traffic-light coalition as the "Svetofor coalition," using the Russian word for traffic light that no English or German writer would employ1

1

5

5

. Another profile described a U.S.-focused Telegram channel as offering "a totally unhackneyed perspective on hazzy"2

2

. ChatGPT was used to strip Russian linguistic markers and generate promotional content, while the plagiarized academic papers and sovereignty index were created without OpenAI's models1

1

.

Limited Reach But Sophisticated Infrastructure Raises Concerns

The influence operation achieved limited immediate impact, with typical social media posts receiving low view counts and official International Burke Institute accounts attracting minimal subscribers3

3

. Telegram channels garnered between 10,000 and 20,000 followers each1

1

5

5

. OpenAI placed this campaign at the lower end of Category Three on the Brookings Breakout Scale, which measures real-time impact and spread of misinformation1

1

. "The significance of the operation lies less in the audience it reached, however, than in the infrastructure it had built," OpenAI stated4

4

. The elaborate construction distinguished this Russian disinformation campaign from other Russia-linked operations OpenAI has disrupted2

2

.

AI Misuse Patterns Emerge Across Multiple Threat Reports

This marks OpenAI's sixth public threat report since early 2024, during which its Intelligence & Investigations team disrupted more than 40 malicious networks1

1

. Similar patterns recur across these threat reports, with operators integrating AI into existing workflows for efficiency and distribution rather than content generation1

1

. In June, OpenAI banned China-linked accounts behind the "Data Center Bandwagon" campaign that amplified U.S. data center electricity backlash1

1

. In February, the company shut down ChatGPT accounts connected to Rybar, a pro-Russia media company described by the UK government as "partially co-ordinated by the Presidential Administration" and funded by Russian state corporation Rostec4

4

. OpenAI noted that AI misuse for spread misinformation on ChatGPT can paradoxically aid detection, as "supporting use of AI can lead to the broader operation being exposed"3

3

.