How OpenAI Caught a Russian Influence Play

It was a 'think tank' with a website, a list of experts, and a sovereign index rating countries. But under the academic veneer, OpenAI saw the workings of a Russian influence operation — and it moved to take it down.

In a report published on its site, the company says it disrupted a social media influence campaign believed to originate in Russia. The tool of choice this time was ChatGPT, which was used to build and promote the International Burke Institute (IBI), a fabricated academic group designed to make Russia look good at the expense of Western nations.

Operating out of an address in Israel, the IBI was crafted to appear legitimate. But a closer look by OpenAI revealed a telltale Slavic linguistic slip, a questionable list of experts, and an index that seemed less than rigorous, to say the least.

The 'think tank' that wasn't

At the center of the campaign was the IBI, registered to an address in Israel, with some evidence suggesting real individuals in Israel represented and promoted the page. However, the core mission was to paint Russia in a favorable light compared to its Western adversaries, particularly those supportive of Ukraine.

The IBI website advertised the 'Burke Sovereignty Index,' which purported to measure how well countries perform across political, economic, technological, informational, cultural, cognitive, and military factors. But according to OpenAI, the Index’s real purpose was to make Western countries look worse than Russia.

OpenAI's analysis points out inconsistencies in the Index's methodology. For instance, The Vatican City, which covers less than half a square kilometer, scored above Spain in its average of scores. Russia, meanwhile, received the highest military score in the Index.

The 'experts' listed

The IBI website boasted a list of experts whose affiliation to the IBI was not referenced. Some are legitimate pioneers in their fields, such as Francis Fukuyama and Noam Chomsky. Others are former high-ranking members of the US government, including Mike Pompeo and Joseph Nye.

However, there are also listings for experts who passed away before the IBI was even founded, such as Shlomo Avineri and Jiang Ping. This glaring anachronism is one of the clearest signs the group was fabricated.

Stolen research, misattributed work

Much of the academic work cited on the IBI website is legitimate, but OpenAI said it has been stolen from its actual authors and misattributed to different authors to hide the true source. This tactic was used to give the IBI website credibility and make it appear authentic to casual visitors.

The operation was careful to cover its tracks. The ChatGPT account operators included prompts that instructed the AI to hide any linguistic clues that the operators used Russian language prompting.

Despite those precautions, a slip slipped through. One inclusion on the IBI website referred to Germany’s traffic light coalition as the 'Svetofor coalition.' Svetofor is the word for 'traffic light' in numerous Slavic languages, including Russian, indicating that drafts were written in Slavic languages before being translated.

The reach of the operation

OpenAI detailed that some of the ChatGPT accounts were used to promote the IBI across X, LinkedIn, Facebook, Substack, and Telegram, using AI-generated imagery and captions. Other accounts were used to automatically engage with comments by real users on Substack.

The operators regularly requested performance summaries of these pages in Russian, and used ChatGPT to generate matching profile pictures for some accounts, OpenAI said.

Impact and dedications

OpenAI noted that the overall impact of the IBI and its promotion was fairly limited. Some Telegram channels garnered followings between ten to twenty thousand, showing a limited breakout to authentic audiences.

But the main notable aspect, according to OpenAI, is the level of dedication behind the underlying infrastructure. The site is designed to look authentic, uses authentic research (even if misattributed), and presents itself as a collection of experts. To a layman, a cursory glance would give the operation a level of authenticity not seen in other social influence campaigns.

Why this matters for you

This is not just a story about one fake think tank. It highlights a growing tactic in influence operations: using AI tools to scale up the creation of seemingly credible content.

For businesses, researchers, and the public, this raises the stakes for verifying sources. The IBI looked good enough to fool a casual observer, and it used real names and real research to do so. As AI tools become more accessible, the cost of creating such convincing disinformation drops, and the need for critical evaluation of online sources becomes even more pressing. The campaign also underscores the ongoing challenge tech companies face in policing their platforms while balancing user freedom.