OpenAI Disrupts Iranian Operation Using AI to Impersonate Journalists, Spread Disinformation

OpenAI Disrupts Iranian Operation Using AI to Impersonate Journalists, Spread Disinformation

OpenAI has disrupted two elaborate, long-running influence operations — one Iranian, the other Russian — that used its AI models along with traditional techniques and technologies to spread anti-American and anti-Israeli propaganda worldwide.

The company reported that operatives tied to the Iranian regime fabricated identities for at least seven Western journalists, who pitched false stories to small and midsize online news outlets with substantial readerships around the world. They repeatedly got those stories published, using the fictitious personas to “to launder geopolitical, conflict-related messaging into their target audiences.” The Iranian operation, which OpenAI has dubbed “Bogus Bylines,” began approximately two and a half years ago.

The company determined the articles were AI-generated because users sent instructions to the OpenAI chatbot in Persian.

According to the report:

As well as long-form articles, the Iranian operation generated batches of social media comments, generally on topics related to the US-Iran war. As well as controlling the “think tank,” the Russian operation created fake “leaked” documents and audio scripts, some of which we identified being spread online. Both operations also made heavy use of AI to draft internal reports (the Russian operation did this more than anything else); in both cases, the actors used questionable or outright deceitful methodologies to exaggerate the operators’ effectiveness.

Using the IO Breakout Scale, which gauges an influence operation’s potential reach on a scale from 1 (lowest) to 6 (highest), the company assigned the Russian operation a rating of 5 and the Iranian operation a 4. The report notes that “both managed to land their content in mainstream media outlets, rather than simply posting it on social media.”

The use of AI helped threat actors “to conduct operations with greater scale, efficiency, linguistic fluency, and editorial ability,” according to OpenAI.

Since July 2025, the seven known invented journalists, Ervin B. Hoskins, Noah Lamington, Sophia Gonzalez, Michael Harrison, Ericka Feusier, Jenny Williams, and Alice Johnson, managed to publish 100 fictitious articles in global media outlets.

OpenAI reported that after they revealed “Alice Donovan” and “PeaceData” as Iranian operatives posing as journalists, they went silent.

Just when social media seems it is drowning in antisemitism and fake news, OpenAI reveals it is even worse.

Iran ran an AI-powered influence op (“Bogus Bylines”) that invented at least seven fake Western “journalists” — Ervin B. Hoskins, Noah Lamington, Sophia Gonzalez, Michael…

— Trisha Posner (@trishaposner) October 13, 2026

The Washington Post reported that at least 20 media sites have published the fictitious content. Op-eds with Ervin Hoskins byline have been published on sites such as the Daily Kos and Middle East Monitor, “which together have more than 3 million followers on Facebook,” according to the article.

Both sites told the Post that after hearing from OpenAI, they had removed the articles published under the bogus bylines and were looking at ways to prevent it from happening in the future.

The Port St. Joe Star, a small local news outlet based in Port St. Joe, Florida, ran an opinion column “written by” Hoskins on September 22, complete with his AI-generated headshot. The piece is titled, “The cost of American power in the Iran War.”

Reached by the Post for comment, the outlet’s publisher David Adlerstein, responded by email:

He wrote that “Hoskins’ column had caught his eye after it was pitched to him by a PeaceVoice, a small liberal think tank in Oregon, that had been sending the Star content for ‘years.'”

Because we’re in a very red part of Florida, I usually don’t run PeaceVoice columns, but this one struck me as thought-provoking and not a strident attack on Mr. Trump, which would be problematic for much of our readership.

Adlerstein added that because of his “existing relationship with PeaceVoice,” he did not find it necessary “to verify the origin or the column or the author’s identity.”

The Post also reached out to Tom Hastings, the founding director of PeaceVoice, who said “the organization has beefed up its existing no-AI policy after OpenAI contacted him.” Good idea.

AI has made deception more sophisticated, and editors must recognize that the playing field has changed. These schemes succeeded because outlets published work by unfamiliar contributors without verifying their identities. Once those articles appeared on established news sites, Iranian and Russian propaganda acquired the credibility of independent journalism — exactly as its creators intended.

But how many people read those articles before they were removed? How many accepted their claims as fact? How many votes were influenced by these lies?

Editors need to wake up. Their credibility is a valuable asset, and hostile regimes are counting on them to hand it over.

[Featured image made with ChatGPT]

Donations tax deductible
to the full extent allowed by law.