Artificial intelligence giant OpenAI has announced the banning of accounts tied to Russian and Iranian influence operations. The company stated that these groups leveraged its ChatGPT platform to create and disseminate AI-powered fake news campaigns targeting audiences across Latin America and international news platforms.
According to OpenAI's October 8 report, the operations combined AI-generated material with traditional influence tactics, including the use of fabricated identities, misleading reports, and politically charged articles. While some activities involved posting comments on social media, a key tactic was attempting to lend credibility to their content by placing articles in existing online publications.
Russian Operation: "Dark Clark"
OpenAI identified and banned a cluster of ChatGPT accounts originating from Russia, labeling the operation "Dark Clark." This campaign primarily targeted countries in Latin America, with much of the content aimed at undermining Ukraine's reputation. There were also efforts focused on domestic politics in Argentina and Bolivia.
The operators behind "Dark Clark" reportedly controlled 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 regional relations. OpenAI's evidence suggests that local staff working for SRC may have been unaware of the Russian backing.
The campaign further involved the creation of fake documents, audio scripts, and misleading material, including content related to claims about Ukraine's honorary consul in Panama. OpenAI rated "Dark Clark" as Category 5 on its six-point influence-operation scale, indicating the highest level of disruption the company has reported since it began tracking such activities.
Iranian Operation: "Bogus Bylines"
Concurrently, OpenAI banned Iran-origin accounts linked to an operation it named "Bogus Bylines." This campaign utilized seven fabricated journalist identities to pitch geopolitical articles, largely focusing on the US-Iran conflict, to various online publications.
OpenAI identified nearly 100 articles published or syndicated under these operation's bylines across approximately a dozen outlets. While the operators also generated numerous political comments, the company found little evidence that these attracted substantial engagement. Internal reports from the operation appeared to inflate their campaign's effectiveness through misleading calculations.
OpenAI has shared information regarding both operations with relevant authorities. The findings underscore how AI tools can streamline and amplify covert influence campaigns, while established publications can inadvertently provide a wider reach for misleading content.