Iran has a long history of using covert media to shape foreign opinion and the most recent discovery by OpenAI reveals how Iranian actors have leveraged advanced Artificial Intelligence tools to launch a media influence operation in the United States. According to a recent OpenAI statement, the company blocked two information-influence operations attributed to Iran and Russia.

The report highlights that actors – likely originating in Iran – used OpenAI models and fake personas to publish critical news articles about the U.S.-Iran conflict across multiple American media outlets. After the articles were published, they used ChatGPT to generate comments for social media to amplify the reach. These comments were focused on themes such as missile attacks, ceasfires, and diplomatic maneuvers.

OpenAI dubbed this operation “Bogus Bylines,” noting that approximately 100 fake articles centered on the U.S.-Iran conflict were published. Interestingly, the operation employed seven fake journalist identities to send these articles to online media outlets. To obfuscate their location, the actors used VPNs to access ChatGPT. Between July 2023 and October 2026, these seven “fake journalists” published 100 articles across various global media.

OpenAI also revealed that several ChatGPT accounts used specifically for editing long articles and generating social media comments in Persian were banned. One example of these articles was titled: “The Iran War and the Collapse of Western Moral Credibility.” Using the IO Breakout Scale, on a scale of 1 (lowest) to 6 (highest), OpenAI rated the intensity of this influence operation breakout at a Level 4.

As we previously reported, the role of AI and agentic AI in proganda battles has become evident globally, and this news demonstrates the growing threat of AI-driven influence. One of the most famous examples of AI-generated content in this ongoing war has been the viral “Lego videos” which mocked U.S. leaders.

While the outlets that published these articles weren’t all mainstream giants, many had significant reach; for instance, Middle East Monitor has nearly two million Facebook followers. A representative from the outlet stated that the articles had been deleted and they were now reviewing their process for submissions.

Whilst OpenAI has not attributed the activity to Iran they added: “The overall activity appeared consistent with a commerical actor running a for-hire campaign, but we have been unable to identify the particular actor involved.”