The Newspaper Was Real. The Byline Wasn’t.

The article can be in a real newspaper and the person supposedly writing it can still be made up. That’s the uncomfortable part of an investigation OpenAI disclosed on October 8, 2026.

OpenAI says it banned ChatGPT accounts connected to an influence operation originating in Iran. According to the company, people used ChatGPT to refine articles, tailor pitches to publications, and support seven fictitious journalist identities. OpenAI identified almost 100 articles published or syndicated under those names between July 2025 and October 2026. October 8 is when OpenAI disclosed the operation, not when all those pieces appeared.

This wasn’t just somebody posting under a fake name on an account with eleven followers. The Washington Post reported that one opinion column ran in Florida’s Port St. Joe Star. Its publisher told the Post he hadn’t checked the author’s identity because the column arrived through a familiar contributor organization. The Post also reported that Middle East Monitor and Daily Kos removed articles under the identified bylines.

OpenAI says it banned the associated accounts. That doesn’t mean every published copy or repost has vanished.

The trick here wasn’t that AI produced prose with some telltale robotic smell. OpenAI describes AI being used to help prepare and place material while people presented invented journalists as real ones. Staring at the sentences and asking whether they “sound like ChatGPT” misses the part that worked: a real publication made an unfamiliar name look more credible.

So what can you do when an article makes a claim you might repeat, share, or act on? First, see what kind of piece it is. A staff-reported story, an opinion column, and a community post have different paths onto a website. The Post noted that Daily Kos’s community section did not receive staff review before publication. The logo at the top of a page doesn’t tell you who checked what.

Then look beyond the byline. Can you find verifiable information about the author? More importantly, has another outlet independently reported the consequential claim, or is everyone linking back to the same article? None of these checks is a foolproof fake-author detector. An unfamiliar writer might simply be unfamiliar.

And a fake identity does not, by itself, prove every sentence in an article is false. Criticism of a government isn’t evidence of an influence operation, either. The deception OpenAI identified concerns who was behind these submissions. The factual claims in any particular piece still need to be judged on their own merits. OpenAI did not identify the particular actor behind the campaign it described as originating in Iran.

You don’t need an AI-detection subscription or a special browser button for this. If a piece matters enough to change your mind, check what kind of article you’re reading and whether its important claims stand up somewhere else. A real website is a good start. It is not a background check.

Information checked: October 11, 2026.