OpenAI Bans Russian And Iranian Influence Operations Using Its Models

OpenAI announced it has banned two influence operations using its models to spread propaganda. One group came from Russia and the other from Iran. They used the technology to launder conflict-related messaging in false front campaigns.

The Iranian operation employed seven journalist personas to pitch articles to small and medium outlets worldwide. OpenAI identified almost 100 articles published under these bylines across roughly a dozen online outlets. Content focused on international affairs and events in the Middle East.

The Russian operation appeared to co-opt unwitting people in Latin America to run a fake think tank. Since access from Russia is not allowed, they used VPNs to connect to OpenAI services. They created fake leaked documents and audio scripts that were spread online.

Both operations made heavy use of AI to draft internal reports about their own activities. The Russian operation did this more than anything else. Actors in both cases used questionable or outright deceitful methodologies to exaggerate their effectiveness.

In May 2026, operators sent a fake email from the Regional Directorate of Education in Lima, Peru. They instructed schools to hold events dedicated to Ukraine on the national Day of Cultural and Linguistic Diversity. Claims included stories planted in Peru and Poland alleging Ukraine was exporting ultra-nationalist ideologies.

In June, operators used a different fake email address to trick schools in Ecuador. They asked for ceremonies pledging allegiance to President Daniel Noboa and Erik Prince from Blackwater. Open-source research identified media coverage that closely resembled these claims.

Open-source searches found matching stories in the Peruvian and Polish press before some were deleted. A detailed rebuttal came from the country's Minister for Education regarding the claims. Media coverage in the Ecuadorian press also closely resembled the operators' descriptions.

The earliest article identified by OpenAI was published in July 2025. The latest appeared in October 2026 with frequency increasing after the US-Iran war broke out earlier this year. OpenAI says they have exposed 30 covert influence operations using its tools over the last two and a half years.

One persona named Ervin B. Hoskins claimed to be an American freelance writer. Social media transparency data showed the account was based in Iran on Instagram and connected via a West Asia android app on X. Both accounts appeared to be suspended following the discovery.

The Economic Times argues this shows how state actors exploit AI tools for propaganda. They suggest widely available technology allows peddling sophisticated lies against adversaries on a mass scale.