Iranian operatives got ChatGPT-written fake articles placed inside real U.S. news outlets.
According to a report published October 9, 2026, people linked to an Iranian influence campaign used OpenAI's ChatGPT to write articles that were then placed in real U.S. news publications. The reporting doesn't name which outlets ran the material, how many pieces got through, or over what timeframe. It also doesn't explain how the articles were caught or whether any outlets have since corrected or retracted them. That leaves plenty of unanswered questions about scale and which newsrooms were actually targeted.
Getting an AI-drafted piece under a real byline carries more weight than posting it to a shady blog or a bot account, because readers extend real outlets a trust that fringe sites don't get for free. That also makes this harder for editors to catch than a bot farm - the usual tells, like broken English or no verifiable history, are exactly what a capable chatbot can paper over.
Don't expect much more detail soon. Specificity here would mostly hand future operators a blueprint, so vagueness is probably doing double duty as both caution and cover.