How did Iranian operatives use AI to publish fake news in American media?
Ervin B. Hoskins, an American freelance journalist, has written dozens of pieces in recent months critical of the United States’ war against Iran.
Published in small American outlets, his provocative pieces portrayed the United States as the aggressor in the now seven-month war, Israel as an untrustworthy ally and Iran as the weaker party that refuses to surrender to its opponent.
But it turns out Hoskins is not an American journalist. In fact, he does not even exist.
Hoskins was one of seven fake journalist personas, posing as Westerners, that Iranian operatives used to publish pieces in media outlets around the world, including in the United States.
This was announced by the company OpenAI, which in a report published on October 8 said it had disrupted an influence operation from Iran that used ChatGPT to secretly spread propaganda against the United States.
“We identified nearly 100 articles published or republished under the bylines of this operation’s authors on about a dozen online platforms around the world,” OpenAI said.
“These were small and medium-sized outlets, generally focused on international affairs, geopolitics and events in the Middle East.”
The company said it had banned the ChatGPT accounts that used the model to generate articles and story pitches, as well as to create comments on social media.
The Iranian operation has highlighted how artificial intelligence systems, now widely accessible, can be misused to spread ever more sophisticated propaganda against a country’s adversaries, on a scale unprecedented until now.
Provocative texts
Among the pieces attributed to Hoskins was one titled “Jared Kushner and the privatization of America’s Iran strategy,” published on June 12 by PeaceVoice, a service for distributing opinions by public intellectuals and media content, run by the Oregon Peace Institute.
Over the following days, the same article was republished under similar headlines by CounterPunch, an American left-wing online magazine, and LA Progressive, which describes itself as an “independent publisher, committed to justice,” based in Los Angeles.
The personal financial interests of Jared Kushner, the son-in-law of US President Donald Trump and his special envoy, his diplomatic activities and possible conflicts of interest, have been covered by major American media, including an investigation by The New Yorker magazine.
However, the piece published by PeaceVoice, CounterPunch and LA Progressive, although it notes that Kushner has defended his business deals and that, according to him, the allegations do not prove corruption, is largely one-sided and devotes little space to his defence.
The piece links Kushner’s financial ties to possible conflicts of interest and suggests that they influence US policy toward Iran, but presents no clear evidence for this claim.
It also overlooks other reasons behind US policy toward Iran, including concerns about Tehran’s nuclear programme and the US’s longstanding ties with key allies such as Israel and Saudi Arabia.
The piece also uses terms such as “privatization,” which make its one-sided editorial stance apparent.
“Cognitive warfare”
Six other fake journalist personas were used by Iranian operatives. Each had social media profiles and AI-generated images, and amplified the others’ content.
Artificial intelligence is also being used to pitch long articles to small and medium-sized outlets around the world. The fake journalists’ articles appeared not only in left-leaning American media, but also in liberal-leaning publications.
The Iranian operation has added to concerns that governments, including Iran’s, are using artificial intelligence models to spread state propaganda and influence public opinion.
OpenAI said the Iranian operation reached category four on its influence operations (IO) escalation scale, which assesses the potential impact of a disinformation campaign. Category six is the highest level on this scale.
“Artificial intelligence can give such covert operations greater reach, greater efficiency, linguistic fluency and better editorial capabilities,” OpenAI said.
“Operators can exploit these advantages to deceive unsuspecting victims, such as employees or editors, and to place their content before audiences who have no idea who is behind it or what their motives are.”
OpenAI is not the only artificial intelligence company involved in identifying such operations.
Last month, the company Anthropic said Iranian state institutions were using artificial intelligence models, such as its Claude app and Google’s Gemini, to wage “cognitive warfare” campaigns and promote narratives created by the powerful Islamic Revolutionary Guard Corps, the elite branch of Iran’s armed forces, which plays an important role in politics and the economy.
Anthropic said it had also disrupted operations by state actors, including Iran, that used Claude to “build, carry out and otherwise facilitate surveillance operations” targeting ethnic minorities and dissidents.
“Artificial intelligence has made it much easier to produce propaganda in many languages, create fake identities and discredit critics,” Pooyesh Azizeddin, a human rights and technology researcher who works on internet censorship, surveillance and digital freedom in Iran, told Radio Free Europe’s Radio Farda.
“On the other hand, authoritarian governments often use the fight against disinformation as a pretext for censorship, expanding surveillance and even for pressuring for the disclosure of users’ identities,” he added.



Komentet
Bëhu i pari që komenton!
Lini një Koment të Ri
Për t'u përgjigjur një komenti specifik, kliko butonin 💬 Përgjigju poshtë atij komenti.