OpenAI bans Russian and Iranian accounts over AI-powered fake news campaigns targeting global audiences

OpenAI has banned ChatGPT accounts linked to Russian and Iranian influence operations that allegedly used AI to support covert campaigns targeting audiences across Latin America and international news platforms. The operations combined AI-generated material with traditional influence tactics, including fabricated identities, misleading reports and politically charged articles. While some activity involved posting comments on social media, the campaigns also sought to make their content appear credible by placing articles in existing publications.

According to OpenAI's October 8 report, the operators used ChatGPT for research, content creation and internal reporting, while sometimes exaggerating the apparent success of their campaigns. One operation went further, allegedly building a research organisation that relied on local staff who may not have known who was behind it.

Must Read: Trump calls people using term ‘Artificial Intelligence’ ‘THE ENEMY’: What his new AI order means

Russian operation used fake research platform and identities

OpenAI said it banned a cluster of Russia-origin ChatGPT accounts involved in an operation it named “Dark Clark”. The activity targeted countries across Latin America, with much of the content apparently aimed at undermining Ukraine's reputation and some efforts focused on domestic politics in Argentina and Bolivia.

The operators appeared to control a purported research platform called the Social Research Center (SRC) through a fabricated persona named “Mia Clark”. The platform’s stated interests included the Indian diaspora in Latin America and relations between India and countries in the region. OpenAI said available evidence suggested that employees working on the ground may have been unaware of the Russian operation.

Must Read: Anthropic tightens Claude AI rules on model abuse, elections, weapons and surveillance

The campaign also involved fake documents, audio scripts and misleading material, including content related to claims about Ukraine's honorary consul in Panama. OpenAI rated the operation Category 5 on its six-point influence-operation scale, its highest rating for a campaign it has disrupted since it began reporting on such activity.

Iranian operation used fake journalist identities

OpenAI also banned Iran-origin accounts linked to “Bogus Bylines”. The operation used seven fabricated journalist identities to pitch geopolitical articles, largely about the US-Iran conflict, to online publications.

OpenAI identified almost 100 articles published or syndicated under the operation's bylines across roughly a dozen outlets. The operators also generated batches of political comments, but the company found little evidence that these attracted substantial engagement. Internal reports appeared to inflate the campaigns' effectiveness through misleading calculations.

OpenAI said it shared information about both operations with relevant authorities. The findings highlight how AI can make covert campaigns easier to organise, while established publications can give misleading content a wider potential reach.