OpenAI says it banned a cluster of ChatGPT accounts originating in Iran after linking them to a campaign that used seven invented journalist identities. The company identified nearly 100 articles published or syndicated under those bylines across more than a dozen online outlets.
The operators used ChatGPT to refine English-language articles about U.S.-Iran geopolitics to fit the submission rules of specific outlets. After approving edits, they asked the model to draft email pitches to editors. The seven bylines presented their authors as Western journalists, often covering politics, human rights or conflict.
OpenAI rated the article-publishing activity Category 4 on a six-level Breakout Scale. Brookings describes Category 4 as a campaign that breaks out of social media and is amplified by mainstream media.
OpenAI could not identify the commercial actor and cautioned that not all published content was generated with its models. The article count reflects the company’s account data and open-source searches; it does not show how many readers were persuaded. The report describes AI-assisted editing and pitches, with operators approving material, rather than autonomous publication.
.jpg){kind=link}