OpenAI said it has banned a cluster of ChatGPT accounts originating in Russia after identifying their use in a covert online influence campaign designed to spread misinformation.
The company said it uncovered the activity while reviewing AI-generated social media posts, a probe that ultimately revealed what it described as a far broader operation. In a blog post published Tuesday, OpenAI said the campaign was built around a website featuring copied and misattributed academic work, a so-called sovereignty index that portrayed Russia favorably, and efforts to conceal the operators’ Russian origins.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
A Broader Playbook For Influence Operations
The case underscores how generative AI is increasingly being used as an enabling tool in information operations, not merely as a content generator. According to OpenAI, the operators prompted ChatGPT in Russian to produce social media comments and posts that were later distributed across Substack, Telegram, X, Facebook and LinkedIn. The same operators also instructed the model to obscure linguistic clues that could reveal their identity or geographic origin.
OpenAI said the campaign centered on an organization called the International Burke Institute, or IBI. The institute’s website, the company said, featured articles copied from real academic sources, sometimes with false attribution. It also included a “sovereignty index” that praised Russia and criticized Western countries.
Why The Infrastructure Mattered More Than The Reach
While OpenAI said the immediate reach of the operation appeared limited, it stressed that the real significance lay in the infrastructure the operators had assembled. The campaign presented an apparently credible institution, complete with named experts, republished research and a proprietary risk index—elements that can be used to manufacture legitimacy at scale.
That distinction matters. In modern influence campaigns, credibility is often the product, not the byproduct. A polished website, a pseudo-academic veneer and coordinated social distribution can make false narratives appear institutional, even when the audience remains relatively small.
OpenAI said the episode illustrates how influence actors can use AI to support broader efforts to obscure authorship, build trust and elevate preferred narratives. It also noted that the use of AI can, paradoxically, expose the operation itself when the supporting infrastructure is scrutinized.
Part Of A Pattern Of Pro-Russia Activity
This is not the first time OpenAI has moved against accounts linked to pro-Russia activity. In February, the company said it shut down ChatGPT accounts associated with Rybar, a pro-Russia media organization that the U.K. government has described as partially coordinated by the Russian Presidential Administration and connected to Russian state interests.
OpenAI said it does not permit access to its models from Russia, but added that the operators in this case used VPNs to mask their location. CNBC has contacted the Russian embassy in London and Russia’s Foreign Ministry for comment.
The broader lesson for businesses, governments and platforms is clear: the next generation of influence campaigns may not rely on volume alone. They will rely on process, polish and the strategic use of AI to create the appearance of authority.
