Iran-based operators used OpenAI’s tools to prepare nearly 100 articles published or redistributed across more than a dozen websites under fictitious journalists’ names, according to findings disclosed by the company and reported by The Washington Post.
OpenAI said it blocked accounts linked to the operation, whose users concealed their locations through VPNs. The operators used its models to create seven Western personas, mainly presented as Americans, and develop articles for submission to English-language publications.
The activity dates back to July 2025. Operators gave ChatGPT instructions in Persian to polish English articles about the US-Iran war, adapt them to publications’ submission requirements and draft messages to editors. They also used the tool to create biographies and support the identities of purported freelance writers.
The articles covered subjects ranging from Israeli politics to US foreign policy. Their arguments were often nuanced and were not uniformly anti-Israel.
One fictitious writer, “Noah Lamington,” examined Israel’s ultra-Orthodox military draft dispute and the stability of Prime Minister Benjamin Netanyahu’s government in International Policy Digest. Another, writing as “Hoskins,” discussed France’s entry ban on Itamar Ben Gvir in Modern Diplomacy. While criticizing Ben Gvir and Washington’s response, that article also expressed support for Israel’s security.
An article in LA Progressive questioned whether Jared Kushner’s business ties in the Gulf influenced US policy toward Iran. Other personas included “Erwin Hawkins,” presented as an American freelance writer and anti-war activist with social media accounts, and “Michael Harrison,” for whom operators requested a complete biography.
Sebastian Gorka: Trump’s strategy to defeat Iran, Hamas and global terrorism | American View
The targeted publications were largely outside the mainstream media, though some have substantial online audiences. London-based Middle East Monitor, which has nearly two million Facebook followers, published at least 21 articles under four bylines that OpenAI identified as fictitious, according to the report. The outlet said it removed articles identified as AI-generated and was reviewing its opinion submission procedures.
Material also appeared in Daily Kos’s community section, where posts are not reviewed by the site’s editors. Publication on established platforms, followed in some cases by social media promotion, helped the fabricated identities reach readers.
The operation extended beyond articles. OpenAI said the network generated social media comments portraying the United States as aggressive, Israel as an unreliable ally and Iran as a victim resisting pressure. Some comments promoted articles the operators had placed themselves.
The network also targeted Iran International, a Persian-language broadcaster critical of Tehran. According to OpenAI, operators used ChatGPT to produce dozens of batches of hostile replies to the outlet’s posts.
An article published under the name “Sofia Gonzalez” in LA Progressive and CounterPunch alleged that Netanyahu pushed Trump into war and accused Iran International of helping create the impression that Iranians welcomed the attacks.