OpenAI announced Thursday that it had shut down a series of accounts linked to separate propaganda efforts originating in Iran and Russia. The Russian campaign, according to the company, achieved the widest audience of any such operation it has identified and disrupted to date.
The two influence operations combined OpenAI’s tools with “more traditional techniques and technologies to support sophisticated ‘false front’ entities,” the company said, using artificial intelligence to covertly advance the political objectives of Tehran and Moscow. The schemes involved generating and distributing fabricated news reports and operating a sham think tank.
“They used those entities to launder geopolitical, conflict-related messaging into their target audiences,” OpenAI wrote in its report. “The Iranian operation included a stable of seven ‘journalist’ personas which it used to pitch long-form articles to small and medium online outlets around the world; the Russian operation appears to have co-opted unwitting people in Latin America to run a ‘think tank’ on the ground.”
The Iranian effort, dubbed “Bogus Bylines,” relied on virtual private networks to obscure the true locations of its operators, who created multiple “American freelance writer” identities. Despite the VPNs, transparency settings exposed the accounts as originating from Iran. Under these false personas, the operatives produced fabricated stories about a U.S.-Iran conflict and flooded social media with comments on geopolitical topics in both English and Persian.
The Russian operation, which OpenAI named “Dark Clark” after a fictitious persona central to the scheme, generated fake “leaked” documents and audio scripts, including some designed to “undermine recruitment for the Ukrainian Armed Forces.” The network also sought to inflame existing tensions between Ukraine and Poland. As part of the plot, it produced reports tied to the “Social Research Center,” a fake think tank that recruited employees across Latin America and deceived institutions throughout the region.
Both operations made extensive use of AI to draft internal reports, with the Russian campaign relying on this tactic more heavily than any other. “In both cases, the actors used questionable or outright deceitful methodologies to exaggerate the operators’ effectiveness,” OpenAI noted.
The disclosure follows a similar report from OpenAI in 2025, which warned that U.S. foreign adversaries were using ChatGPT alongside other AI models to “achieve their aims,” including a Russian cyber operation that posted AI-generated content across multiple platforms.
The latest findings arrive amid declining public confidence in artificial intelligence and in lawmakers’ capacity to oversee its development. A recent Reuters/Ipsos poll found that 57 percent of respondents believe the Trump administration is not taking AI risks seriously enough, while 54 percent said the same of Congress. The administration has generally favored voluntary safeguards over broad federal regulation. President Trump invited leaders from major tech companies—including OpenAI, Nvidia, Anthropic, Meta, and SpaceX—to sign a “morally-binding” self-policing pledge that includes external reviews of their AI systems. Critics remain skeptical that such self-regulation will be effective.
OpenAI said that publicizing AI-based influence operations like those exposed Thursday will make it easier for the company to conduct “further research and disruption” and prevent similar schemes in the future. The company added that its reports have, in some instances, deterred such activity.
