OpenAI says operatives in Iran used ChatGPT to plant about 100 articles in news outlets
OpenAI said on Thursday that operatives in Iran used its models and fake personas to help plant news articles critical of the US war against Iran in multiple US media outlets. According to the company's blog post, about a dozen outlets around the world ran roughly 100 articles credited to seven fake journalists. OpenAI banned the ChatGPT accounts involved. They had been prompting the tool in Farsi to tweak articles and editorial pitches and to generate social media comments.
Some of the AI-assisted pieces carried provocative titles, among them 'Is America’s Iran Policy Being Privatized? The Jared Kushner Question' and 'The Iran War and the Collapse of Western Moral Credibility'. None of the publications that ran the articles appear to be a major mainstream news outlet, CNN reports, but some have sizable online followings. Middle East Monitor, which is based in London, has almost two million followers on Facebook. The Kushner article was published by LA Progressive, which describes itself as an 'independent, justice-driven publication' based in Los Angeles.
After the articles went up, the operatives asked ChatGPT to generate comments they could post on social media purporting to be people reacting to them. OpenAI's blog says the users asked for comments on topics such as missile strikes, ceasefires and diplomatic maneuvering, specifying outputs that portrayed the US as the aggressor, Israel as an unreliable ally of the US, and Iran as a defiant but resilient victim.
Who was behind the operation, and whether they are affiliated with the Iranian government, is unclear, OpenAI said. In its words, the overall activity appeared consistent with a commercial actor running a for-hire influence campaign, but the company is unable to identify the particular actor involved.
Middle East Monitor responded after CNN reached out. Daud Abdullah, Director of the Middle East Monitor, said he would look into the articles OpenAI flagged and 'take the appropriate action'. Another Middle East Monitor representative later said the articles had been taken down and that the publication was reviewing its submissions process. CNN has requested comment from LA Progressive.
Jessica Brandt, who ran the Office of the Director of National Intelligence's now-shuttered Foreign Malign Influence Center from 2023 to 2025, told CNN it is remarkable that the Iranians managed to land their content in real media outlets, expanding the potential reach of their campaign. She said AI is letting foreign adversaries better hide their hand and build more sophisticated false fronts to launder content through authentic channels.
CNN frames the case as the latest example of AI playing a key role in the propaganda battles around the US and Israel's war with Iran. It points to young Iranians using AI to make viral Lego-themed videos mocking US leaders, and to President Donald Trump posting AI-generated videos depicting US strikes on Iran. It also recalls its earlier reporting that, this spring, a false AI-assisted US intelligence report about a purported Chinese ship transporting nuclear components almost led the US military to board the ship.
Key facts
- OpenAI says operatives in Iran used its models and fake personas to help place roughly 100 articles, credited to seven fake journalists, in about a dozen outlets worldwide, including US ones.
- The banned ChatGPT accounts prompted the tool in Farsi to tweak articles and pitches, then generated social media comments posing as readers reacting to the articles.
- OpenAI says the activity looked consistent with a commercial actor running a for-hire influence campaign, but it cannot identify the actor or say whether the Iranian government is involved.
- Only two outlets are named: Middle East Monitor (London, almost two million Facebook followers), which took the articles down and is reviewing its submissions process, and LA Progressive, which had not responded to CNN.
- Former intelligence official Jessica Brandt says AI helps foreign adversaries build more sophisticated false fronts and launder content through authentic channels.
Why it matters
The case moves AI-assisted influence work from social media posting into real publications. Jessica Brandt, who ran the Foreign Malign Influence Center from 2023 to 2025, says landing content in real outlets expands the potential reach of a campaign, and that AI helps adversaries hide their hand behind more sophisticated false fronts. CNN places it in a wider run of AI use around the war, from Iranian Lego-style videos mocking US leaders to AI-generated videos posted by President Trump to a false AI-assisted intelligence report that almost led the US military to board a Chinese ship.
Who it affects
Outlets that accept outside contributions are the direct targets. Middle East Monitor, based in London with almost two million Facebook followers, and LA Progressive, based in Los Angeles, are the two named. Readers in the US are the intended audience: CNN notes that getting anti-war narratives in front of a US audience could count as a successful information operation for whoever was behind it. OpenAI and other AI providers are affected as the platforms whose accounts were used.
How to use it
This is not a product or a tool, so there is nothing to adopt. The practical signal comes from the outlets' response: Middle East Monitor took the articles down and said it was reviewing its submissions process. Editors and platform teams can read the case as an example of what planted, AI-tweaked contributions with fake bylines look like, along with the follow-up social media comments that posed as organic reactions.
How solid is it
The core facts come from OpenAI's own blog post, relayed by CNN, and the figures are approximate: about a dozen outlets, roughly 100 articles, seven fake journalists. CNN confirmed responses from Middle East Monitor, including the takedown. LA Progressive had not responded when CNN published. Two article titles are quoted, and only two outlets are named. CNN's comment that this could count as a successful information operation is its own assessment, not a measured result.
Risks and caveats
Attribution is open. OpenAI says it cannot identify the particular actor and that it is unclear whether the actor is affiliated with the Iranian government; its assessment is that the activity looked consistent with a commercial actor running a for-hire campaign. The source gives no figures on reach, engagement or impact of the articles or comments, and does not say how many of the roughly 100 articles ran in US outlets. It also does not say whether the outlets knew the bylines were fake. The picture rests on one company's account of accounts it banned.
“The overall activity appeared consistent with a commercial actor running a for-hire influence campaign, but we are unable to identify the particular actor involved”
— OpenAI, blog post