OpenAI has banned a cluster of ChatGPT accounts that it said originated in Russia and were used to support a covert online influence campaign, according to CNBC. The company said it uncovered the accounts while investigating AI-generated social media posts, which led it to what OpenAI described as a broader operation. According to CNBC’s account of OpenAI’s Tuesday blog post, the campaign centered on a website that included copied and misattributed academic work, a “sovereignty” index that presented Russia favorably, and steps intended to obscure the operators’ Russian origins. OpenAI said the influence effort appeared limited in immediate audience impact, but more significant for the infrastructure it had assembled. The operators allegedly prompted ChatGPT in Russian to produce social media comments that were posted across Substack, Telegram, X, Facebook and LinkedIn, CNBC reports. OpenAI said the operators also instructed ChatGPT to remove linguistic signals that could reveal they were Russian. A central asset in the operation was an organization called the International Burke Institute, according to CNBC. OpenAI said the institute’s website featured articles copied from real academic writing, in some cases with false attribution, and was promoted through AI-generated posts. The operation also created a “sovereignty” index, OpenAI said, which praised Russia and criticized Western countries. CNBC reports that OpenAI framed the campaign as an example of how influence operators can use generative AI not necessarily as the whole operation, but as a supporting layer for producing content, lending apparent authority and masking the source of favored narratives. OpenAI does not allow access to its models in Russia, CNBC reports, but the company said virtual private networks were used to disguise location. That detail matters because it points to a recurring enforcement problem for model providers: access restrictions can be undermined by basic circumvention tools, leaving detection to usage patterns, content analysis and operational links. The case follows an earlier OpenAI takedown in February, when the company said it had shut ChatGPT accounts connected to Rybar, a pro-Russia media company. CNBC notes that the U.K. government has described Rybar as partly coordinated by Russia’s Presidential Administration, funded by Russian state corporation Rostec, and working with members of Russian intelligence services. CNBC said it contacted the Russian embassy in London and Russia’s Foreign Ministry for comment. Within this cluster, no response from those entities is reported. Who benefits: OpenAI benefits if it can show that misuse of its systems can lead investigators back to broader influence infrastructure. Platforms named in the report may also benefit from clearer indicators of coordinated content tied to the operation. Who's exposed: Influence operators are exposed when their AI usage leaves patterns that connect social posts to a broader campaign. OpenAI remains exposed to scrutiny over how effectively it can enforce geographic restrictions and misuse policies.