Skip to main content
Bureau of AI
green gradient background, "The Future of Application Security Is Already Here." and a read the report button.
IncidentPeople used AI as a toolSeverity M: moderate, 3/5Confirmed: a company, official or court statement

OpenAI said Iran-linked operatives used ChatGPT to prepare and place anti-US war articles in media outlets

OpenAI said Iran-linked operatives used ChatGPT, prompted in Persian, to revise articles, prepare editorial pitches and generate social media comments for an influence campaign against the US war with Iran. The operation used seven fake journalists, and roughly 100 articles appeared in about a dozen publications, including the LA Progressive and Middle East Monitor. AI-generated comments were then written to portray the US as the aggressor. OpenAI banned the accounts but could not establish who was behind the operation or whether it was tied to the Iranian government.

What the AI did

People reported to be linked to Iran used ChatGPT, prompting it in Persian, to revise articles, prepare editorial pitches and generate social media comments for an influence campaign against the US war with Iran. The articles ran under the names of seven fake journalists.

First reported October 8, 2026 · Added to the register October 17, 2026 · 6 sources

Developer
OpenAI
Deployer
Iran-linked operatives
Model
ChatGPT
Affected
LA Progressive, Middle East Monitor, Middle East Monitor (at least 21 articles), Port St. Joe Star, Daily Kos, CounterPunch and LA Progressive, PeaceVoice, Iran International
When it happened
Not stated in the sources
First reported
October 8, 2026

What this means for you

Could this affect you?

Yes

If you publish contributed articles or host comments. Smaller publications that accept contributed articles and pitches, and their readers, received AI-assisted propaganda presented as independent journalism under fake bylines, with roughly 100 articles appearing in about a dozen publications.

What to check

  • Verify the identities of contributors before publishing their work.
  • Check pitches and op-eds for signs of coordinated or AI-assisted submissions.
  • Watch for comment campaigns that follow newly published articles.

Areas of your AI programme this touches

Content and platform moderationDeepfakes and impersonation

Timeline

  1. October 8, 2026First reported
  2. October 9, 2026OpenAI said the earliest article it identified from the Iranian operation was published in July 2025 and the latest in October 2026, with frequency increasing after the US-Iran war broke out. One persona, Ervin B. Hoskins, billed as 'an American freelance writer,' had X and Instagram accounts whose transparency information showed a 'West Asia android app' and an Iran base respectively, and both accounts appeared to be suspended.[2]
  3. October 9, 2026OpenAI called the operation "Bogus Bylines" and rated its article-placement campaign Category 4 on the six-level IO Breakout Scale. Middle East Monitor, which The Washington Post reported had published at least 21 articles under four of the fictitious names, and Daily Kos removed material linked to the network after being alerted, and PeaceVoice strengthened its existing prohibition on AI-assisted submissions after OpenAI contacted it.[3]
  4. October 11, 2026RFE/RL reported that an article attributed to the fake persona Ervin B. Hoskins, titled 'Jared Kushner and the Privatization of America's Iran Strategy,' was published by PeaceVoice on June 12 and over the next few days was republished under similar headlines by CounterPunch and LA Progressive.[4]
  5. October 11, 2026The Washington Post reported that the Port St. Joe Star, a small Florida newspaper, published an opinion article attributed to the fictitious journalist Ervin Hoskins that had been submitted through PeaceVoice. The network also targeted Iran International, a Persian-language news organization critical of the Iranian regime, generating more than two dozen batches of hostile comments intended for its social media posts.[5]
  6. October 12, 2026A Middle East Monitor representative told CNN that the articles identified by OpenAI had been taken down and that the publication was reviewing its submissions process.[6]
Every fact and its source (12)
  1. OpenAI disclosure date2026-10-08
    “OpenAI said on October 8”[1]
  2. Earliest article publishedJuly 2025
    “The earliest article identified by OpenAI was published in July 2025 and the latest in October 2026”[2]
  3. Number of articles publishedRoughly 100
    “The articles, numbering roughly 100”[1]
  4. AttributionOpenAI could not establish who was behind it
    “OpenAI said it could not establish who was behind the operation”[1]
  5. Publisher responseMiddle East Monitor director to investigate
    “Its director, Daud Abdullah, told CNN he would investigate the flagged articles and take appropriate action.”[1]
  6. How the AI was usedPrompted in Persian to revise articles, pitches and comments
    “The accounts had prompted the AI tool in Persian to revise articles, prepare editorial pitches and generate social media comments”[1]
  7. Fake journalists and outletsSeven fake journalists, about a dozen publications
    “the operation involved seven fake journalists whose articles appeared in about a dozen publications worldwide”[1]
  8. Persona's social accounts statusAppeared suspended
    “Both accounts appeared to be suspended.”[2]
  9. Named fake personaErvin B. Hoskins
    “One of the personas named Ervin B. Hoskins, whose bio claimed to be "an American freelance writer,"”[2]
  10. OpenAI impact ratingCategory 4 on six-level IO Breakout Scale
    “OpenAI assessed the article-placement campaign as Category 4 on the six-level IO Breakout Scale”[3]
  11. Operation nameBogus Bylines
    “The operation, which OpenAI called "Bogus Bylines,"”[3]
  12. Outlet responseArticles taken down; submissions process under review
    “the articles had been taken down and said the publication was reviewing its submissions process”[6]

Sources

  1. Iranian operatives used AI to plant fake stories in US media, generate anti-US war narratives: OpenAI reveals
    hindustantimes.com · October 9, 2026
  2. Iran and Russia used ChatGPT for propaganda by 'fake journalists', OpenAI says
    euronews.com · October 9, 2026
  3. OpenAI bans Iran-origin network behind planted articles
    journalismpakistan.com · October 9, 2026
  4. How Iranian Operatives Used AI To Publish Fake Stories In US Media
    eurasiareview.com · October 11, 2026
  5. IRAN EXPOSED: Fake Journalists Used ChatGPT To Plant Nearly 100 Anti-Israel And Anti-American Articles In Western Media
    theyeshivaworld.com · October 11, 2026
  6. Open AI says Iranian regime used ChatGPT to create fake articles
    voanews.com · October 12, 2026

How this record is classified. Severity M (3/5): harm to individuals, or a policy breach with limited reach. OECD level: incident, an event in which an AI system led to actual harm. Evidence: Confirmed, meaning a company, official or court statement.

The facts in At a glance and in Every fact and its source are quoted from the sources listed. The summary, What the AI did and What this means for you are written from those sources and checked against them automatically before publication. Records reflect what has been disclosed, not everything that has happened. Incidents are found through news feeds and the GDELT Project.

Promotional banner for the Penetration Report Template Kit