Iranian Network Used AI to Plant Critical Articles in Global Media
Translated & summarized from Now 14 by baba
An Iranian-based network used artificial intelligence, including ChatGPT, to publish around 100 articles critical of U.S. and Israeli policy in about a dozen global media outlets, according to OpenAI. The operation created seven fictional journalists and also generated fake social media comments to simulate public discussion. OpenAI has blocked the accounts involved, though the exact operator remains unidentified, and experts note AI's role in enabling such sophisticated influence campaigns.
The story in 5 lines · by baba
- An Iranian network used AI to publish about 100 critical articles in global media via seven fictional journalists.
- OpenAI revealed the operation, which utilized ChatGPT to refine articles and generate fake social media comments.
- The articles criticized U.S. policy and the war against Iran, with some targeting figures like Jared Kushner.
- While the operator is unidentified, it appeared to be a commercial influence campaign, not a direct state operation.
- AI enables foreign adversaries to better conceal their activities and create sophisticated false fronts, experts say.
OpenAI has revealed an influence operation originating from Iran that utilized artificial intelligence to disseminate articles critical of U.S. and Israeli actions. The operation involved creating seven fictional journalists who managed to publish approximately 100 articles in about a dozen media outlets worldwide. These articles often criticized U.S. policy and the war against Iran. OpenAI stated that the operators used ChatGPT to refine articles, prepare submission proposals, and generate fake user comments for social media. Following the exposure, OpenAI blocked the accounts involved, though the specific entity behind the operation remains unidentified, with OpenAI noting it appeared to be a commercial entity running a paid influence campaign rather than a direct state operation.
According to OpenAI's findings, the content was crafted in Persian, with requests made to ChatGPT to improve phrasing, tailor submission proposals, and generate additional content promoting critical messages against U.S. policy. Some of the identified article titles included "Is U.S. Policy Toward Iran Shifting to Private Hands? The Jared Kushner Question" and "The War in Iran and the Collapse of the West's Moral Credibility." While the outlets that published these articles were not among major U.S. news organizations, some had significant social media followings, such as Middle East Monitor, based in London, which has nearly two million Facebook followers.
The operation extended beyond article publication. The network operators also used ChatGPT to generate social media comments that appeared to be from individuals reacting to news about missile strikes, ceasefires, and diplomatic moves. These comments aimed to portray the U.S. as the aggressor, Israel as an unreliable ally, and Iran as a determined victim. This combination of published articles and social media reactions was intended to create the impression of an authentic public discussion supporting the disseminated messages.
Jessica Brandt, former head of the U.S. Center for Countering Foreign Malign Influence, commented to CNN that the ability to place content in real media outlets significantly expanded the campaign's reach. She highlighted that AI allows foreign adversaries to better conceal their tracks, build more sophisticated false fronts, and launder their content through authentic channels.