Slashdot has reported:

OpenAI says it recently banned two covert influence operations, one linked to Russia and the other to Iran, that used its AI models in campaigns to spread geopolitical and conflict-related messages. The operations relied on false fronts and placed some material in mainstream media outlets, rather than limiting their activity to social media. Not all of the material they circulated was generated with OpenAI models.

The Iranian operation used seven purported journalist personas to pitch long-form articles to small and medium-sized online outlets around the world. OpenAI identified almost 100 articles published or syndicated under the operation’s bylines across roughly a dozen outlets, which generally covered international affairs, geopolitics and events in the Middle East. The first article identified appeared in July 2025 and the latest in October 2026, with publication becoming more frequent after the US-Iran war began earlier in 2026.

The operation also produced batches of social media comments, generally about topics connected to the US-Iran war. One persona, Ervin B. Hoskins, described in a biography as “an American freelance writer,” had accounts on technology platforms including Elon Musk’s X and Meta-owned Instagram. Screenshots supplied by OpenAI showed X transparency information linking the account to a “West Asia android app,” while Instagram’s transparency information indicated the account was based in Iran. Both accounts appeared to have been suspended.

The Russian operation appears to have recruited people in Latin America who were unaware of the campaign to run a think tank. Because OpenAI does not allow access to its models from Russia, the operators connected to the services through VPNs, the company said. They produced purportedly leaked documents and audio scripts, some of which OpenAI identified circulating online.

OpenAI said it had uncovered 30 covert influence operations using its tools over the previous two and a half years. In this case, both operations also used AI extensively to draft internal reports, with the Russian group relying on it more for that purpose than for any other. OpenAI said actors in both campaigns used questionable or outright deceptive methods to overstate their effectiveness.

The Russian operators’ internal claims included an alleged effort in May 2026 involving schools in Lima, Peru. They said they created a fake email address posing as the Regional Directorate of Education in Lima and used it to tell schools in the district to hold events about Ukraine on the national Day of Cultural and Linguistic Diversity, May 21. According to the operators, some schools replied to the false address, saying they had held events and supplying photographs.

The operators said they then placed media stories in Peru and Poland about the events, alongside allegations that Ukraine was “exporting” ultra-nationalist ideologies, which they said triggered outrage. Open-source searches found stories consistent with those claims in the Peruvian and Polish press, as well as an English-language publication in Hungary. Some of the articles have since been deleted.

In June, the operators said they used another fake email address to persuade schools in Ecuador to hold a ceremony pledging allegiance to President Daniel Noboa and Erik Prince, the former head of private military contractor Blackwater. They claimed the incident caused public anger and pressured the government to deny the falsehood, spreading it to a national audience. Open-source research found similar coverage in Ecuadorian media and a detailed rebuttal from Ecuador’s Minister for Education.

The operators described other tactics in their internal reports. They said they circulated two separate fakes targeting Ecuador in March; one involved fabricated audio attributed to Ukraine’s consul in Ecuador, who was alleged to have made disparaging remarks about Ecuadorians. OpenAI’s report was described by the Economic Times as the latest illustration of how state actors can use widely available AI tools to spread sophisticated propaganda against adversaries at scale.