OpenAI bans Russian and Iranian ChatGPT influence operations

OpenAI said on 8 October it shut down two influence operations, one from Russia and one from Iran, that used ChatGPT to spread fake journalist personas and anti-Ukraine propaganda. The American AI company banned a cluster of ChatGPT accounts that originated in Russia and Iran and were used to support what it called "false front" entities for geopolitical and conflict-related messaging.washingtonexaminer.com+1

The announcement matters because it shows how state actors can use widely available AI tools to produce sophisticated propaganda at scale. OpenAI disclosed the bans in a safety report on 8 October, as public concern about advanced AI technology grows. The two operations differed in structure: one planted articles through fake bylines, the other ran a front think-tank using people who did not know they were working for a Russian group.france24.com

What OpenAI announced

The Iranian operation involved seven profiles posing as Western journalists, including one purporting to be from the United States. These personas were used to place long-form articles in media outlets around the world. OpenAI said it identified almost 100 articles published or syndicated under the operation's bylines across roughly a dozen online outlets, generally focused on international affairs, geopolitics, and events in the Middle East.

The earliest article OpenAI identified was published in July 2025, and the latest in October 2026. The frequency of reports increased after the US-Iran war broke out earlier in 2026. The operation also generated social media comments on topics related to that war, OpenAI said. The outlets placing the articles were small to medium in size, spread across different countries.

One persona, named Ervin B. Hoskins, described itself as "an American freelance writer" and had accounts on X and Instagram. Screenshots OpenAI provided showed X's transparency information linked the account to a "West Asia android app," while Instagram's transparency data placed the account in Iran. Both accounts appeared to be suspended. The mismatch between the claimed American identity and the Iranian signal is how the operation was traced.

How the fake journalists worked

The Russian operation took a different approach: it co-opted unwitting people in Latin America to run a front group called the Social Research Center, or SRC. The group was used for covert activity aimed at undermining Ukraine's reputation in the region and influencing local politics. OpenAI said the available evidence indicates the SRC's employees in Latin America were not aware they were working for a Russian group.

OpenAI said the SRC appears to have produced a majority of its original content through its co-opted staff, rather than importing material from Russia. The company called it the most complex attempt to run a front identity it has disrupted over the past two and a half years. That complexity came from using real, unwitting contributors whose work carried genuine bylines and local credibility.

Both operations relied on the same underlying mechanism: ChatGPT generated the personas, articles, and social media comments that human operators then placed in real outlets and platforms. The Iranian network needed fake journalists with plausible bios; the Russian network needed a credible institutional wrapper. In each case, the AI tool lowered the cost of producing enough content to sustain a long-running campaign across many outlets.

The Latin American front group

OpenAI's report is the latest example of state actors exploiting widely available AI tools to peddle propaganda against adversaries at mass scale. The company did not say whether the operations persuaded any audiences or changed any political outcomes, only that it identified and banned the accounts behind them. The report was published as a safety disclosure, not as an assessment of the campaigns' effectiveness.

The Iranian operation's output grew more frequent after the US-Iran war began in 2026, tying the campaign to a live conflict. The Russian operation targeted Latin America, a region where Ukraine's reputation affects local politics. Together the two networks show influence efforts aimed at different audiences and different goals, both drawing on the same consumer AI product to produce content at scale.

OpenAI banned the account cluster and suspended the traced social media profiles, but the articles already published remain in the outlets that carried them. The company identified almost 100 articles across roughly a dozen outlets, with the earliest from July 2025 and the latest from October 2026. OpenAI's safety report disclosing the bans was published on 8 October.