OpenAI said on October 8, 2026, that it had banned a cluster of ChatGPT accounts tied to an Iranian-linked influence operation that succeeded in getting its work published by legitimate news outlets.

The campaign, which OpenAI has named "Bogus Bylines," fabricated seven Western journalists and used their invented bylines to push nearly 100 articles critical of the US war against Iran.

How the Bogus Bylines Operation Worked

According to OpenAI, the operators interacted with its models primarily in Persian, using them for three core tasks: drafting long, polished articles, writing editorial pitches, and generating social media comments.

The seven fabricated personas carried Western-sounding names, among them Michael Harrison and Ervin B. Hoskins. Under those bylines, the campaign placed approximately 100 articles across roughly a dozen outlets. One headline OpenAI cited read "The Iran War and the Collapse of Western Moral Credibility."

OpenAI said the operators also asked its models to produce comments casting the US as the aggressor and Israel as untrustworthy.

Activity began in July 2025 and continued into October 2026, when OpenAI moved against the accounts. The company said the operation intensified after the US-Iran conflict escalated, with output peaking during the fighting.

Category 4 Rating, With a Russian Operation at Category 5

OpenAI assesses influence operations on what it calls the IO Breakout Scale, a severity rating the company uses to gauge how widely an operation's reach extends. Bogus Bylines landed at Category 4. A Russian campaign the company also disrupted, known as Dark Clark, received a Category 5 rating—a higher tier indicating a more serious threat.

OpenAI was careful on one key point: who was actually behind Bogus Bylines. The company said the operation appeared commercial in nature, and its exact affiliations, including any potential government links, remain unconfirmed.

Background: OpenAI's Running File on Iran

This is not OpenAI's first encounter with Iranian-linked covert activity. The company has publicly reported on such operations since 2024, with earlier Iranian campaigns repeatedly leaning on AI to generate content and steer geopolitical narratives.

What sets Bogus Bylines apart is where the content ended up. Previous campaigns largely stayed confined to social platforms; this one managed to place articles in legitimate outlets.

Implications for Media, Platforms, and Readers

Bogus Bylines exploited the contributor submission system—the channels many publications keep open for outside op-eds and commentary. The fact that seven invented writers could place roughly 100 articles across about a dozen outlets suggests that contributor vetting at many publications is thinner than readers might assume. It is also a reminder that a byline in a legitimate outlet is not, by itself, proof that the writer exists.

The episode also highlights OpenAI's dual role: its models helped produce the content, but the company also detected the activity, shut down the accounts, and published the details.

If the operation was run by a commercial contractor rather than a state agency, it fits a model in which influence campaigns are outsourced—making them harder to sanction and easier to deny.

What to watch next: whether the outlets that ran the Bogus Bylines articles disclose or retract them, whether OpenAI or others firm up who was behind the campaign, and whether other AI providers report similar activity on their own platforms.