OpenAI users in Russia, Iran used models to push political agenda online, in media

OpenAI Bans Russia- and Iran-Linked Influence Operations

Provpnadvice.com – OpenAI users in Russia Iran were tied to separate covert influence campaigns that used AI tools alongside familiar online deception tactics, according to the company. OpenAI said it removed the accounts involved after identifying efforts to spread political messaging through false identities, fabricated organizations and misleading content.

The activity focused on geopolitical conflicts, foreign policy and Ukraine. While generative AI helped operators draft material and support their campaigns, the operations also relied on traditional methods such as fake personas, social-media distribution and deceptive institutions designed to appear credible.

Iran-linked campaign used fake journalist identities

One Iranian operation, known as “Bogus Bylines,” created seven purported journalists who presented themselves as American freelance writers. The personas pitched long-form articles to small and medium-sized online outlets around the world and shared political commentary in English and Persian on social platforms.

OpenAI said account transparency features linked the activity to Iran, despite attempts to obscure the operators’ locations through virtual private networks. The network produced fabricated articles about the U.S. war with the Islamic Republic and sought to amplify wider geopolitical discussions through coordinated comments.

“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,” OpenAI said.

Russian network built a false think tank

A separate Russia-linked operation, called “Dark Clark,” operated through a fabricated organization known as the Social Research Center. The purported think tank recruited people in Latin America and misled regional institutions, giving the campaign an appearance of legitimacy beyond its online presence.

The operation created reports attributed to the false organization, fabricated leaked documents and scripts for audio content. Some material was intended to weaken support for recruitment into Ukraine’s armed forces, while other messaging aimed to exploit tensions between Ukraine and Poland.

OpenAI users in Russia Iran reportedly used its models to support these broader efforts, but the company said the Russian campaign made particularly extensive use of AI for drafting internal reports. It also found that the operators used deceptive methods to exaggerate the campaign’s reach and effectiveness.

“Both operations also made heavy use of AI to draft internal reports,” OpenAI said, noting that the Russian operators used the technology for this purpose more than anything else.

Why the findings matter

OpenAI said the false-front entities were meant to move conflict-related and geopolitical narratives into target communities while concealing the people behind them. The cases show how AI can accelerate parts of an influence operation without replacing the human networks, fake institutions and distribution channels that make disinformation campaigns work.

Removing accounts limits access to a specific platform, but misleading material may continue to circulate elsewhere. Readers should be cautious when assessing claims from unfamiliar research groups, online commentators or news outlets, especially when those sources do not provide clear information about their staff, funding or institutional history.

FAQ: Identifying AI-Assisted Influence Campaigns

What should readers check before sharing political content?

Look for transparent authorship, verifiable contact information, a credible publishing history and independent reporting that supports the claim. Be especially careful with sensational documents, anonymous “leaks” and organizations that appear to have little public record.

Did OpenAI users in Russia Iran create all of the campaign content with AI?

No. OpenAI said the campaigns combined AI-assisted drafting with established influence tactics, including fake personas, fabricated institutions and coordinated distribution. AI was one component of a broader effort to make deceptive messaging easier to produce and adapt.

Can a fake think tank or journalist persona be recognized easily?

Not always. Warning signs can include missing biographies, unverifiable affiliations, recently created social accounts, inconsistent writing histories and claims that cannot be confirmed by reliable independent sources.