OpenAI Says It Broke Up Two “False Front” Influence Operations, One Its Most Far-Reaching Yet

OpenAI says it has banned two covert influence operations, one from Russia and one from Iran, that used its models to support “false front” entities: a fake think tank and a set of fake journalists [1]. These were propaganda campaigns that tried to plant their messages in real media outlets rather than relying only on fake social media accounts. OpenAI rates the Russian one the most far-reaching it has disrupted since it began publishing these reports.

The findings below come from OpenAI’s own threat report of 8 October 2026. TSN has not independently verified them. News coverage is labelled reported.

What is a “false front”?

An influence operation tries to shape opinion while hiding who is behind it. A false front is a fake organisation or persona, such as a think tank or a freelance journalist, that gives the operation’s messages a respectable-looking source. OpenAI points out that this is an old technique; what AI changed is that it made “some of the workflows easier” [1].

Russia: “Dark Clark”

OpenAI says Russia-based operators, using VPNs to reach ChatGPT, ran campaigns across Latin America, mostly aimed at “undermining Ukraine’s reputation in the region”, with some aimed at politics in Argentina and Bolivia [1].

The unusual part was a self-described research platform, the Social Research Center. OpenAI says the operators appear to have controlled it through a fake persona called “Mia Clark”, and that “the SRC’s employees in Latin America were not aware that they were working for a Russian group” [1]. OpenAI calls it “the most complex attempt to run a front identity that we’ve disrupted over the past two and a half years” [1][3].

Most of the operators’ ChatGPT use was not making propaganda. It was writing internal reports to an unknown superior [1]. Those reports described, for example, a fake email to schools in Lima, Peru, asking them to hold events referencing the Ukrainian nationalist Stepan Bandera, a figure associated in Poland with wartime atrocities. OpenAI found matching stories in the Peruvian and Polish press, and some articles carried a reaction from a Polish Member of the European Parliament [1].

Rated Category 5. On the IO Breakout Scale, which runs from 1 (lowest) to 6 (highest), OpenAI puts Dark Clark in Category 5, calling it “the first Category 5 operation we’ve disrupted since we began our reporting” [1]. OpenAI adds a caution: the operators often claimed credit for things they had not done, so “their own claims of impact in their internal reporting cannot be taken at face value” [1].

Iran: “Bogus Bylines”

The Iranian operation used seven fake journalist bylines to pitch long articles, mainly about the US-Iran conflict, to small and medium online outlets [1]. OpenAI says it found “almost 100 articles” published under those bylines, across about a dozen outlets (it gives “roughly a dozen” in one place and “over a dozen” in another) [1]. NPR reported that the Middle East Monitor published at least 21 articles by four of the fake authors [2].

The operation also produced batches of social media comments, which OpenAI says drew little engagement, and internal reports that used a misleading method to inflate its apparent reach [1]. OpenAI rates the article work Category 4. It says the activity “appeared consistent with a commercial actor running a for-hire influence campaign”, but it could not identify that actor [1].

Why it matters

OpenAI says this fits a pattern across the 30 covert influence operations it has exposed since early 2024: those that get content into real media outlets “tend to have the highest potential reach and impact” [1]. That shifts some of the defence to editors, who may be receiving polished pitches from writers who do not exist. OpenAI says it has shared information on both cases with the relevant authorities [1].

What this does not prove

• That the operations changed anyone’s mind. The Breakout Scale measures how far content spread, not its effect on opinion [1].
• That everything the Russian operators claimed is true. OpenAI says their impact claims “cannot be taken at face value” [1].
• Who ran the Iranian operation. OpenAI could not identify the actor [1].
• That AI made these operations possible. OpenAI says they closely resembled pre-AI campaigns; AI made some tasks easier [1].
• Independent confirmation. These are OpenAI’s own findings and ratings.

The Bottom Line

OpenAI’s latest report describes two propaganda operations that hid behind a fake think tank and fake journalists, and used ChatGPT mainly for paperwork and polishing. The Russian one, by OpenAI’s rating, reached further than any it has disrupted before. The practical lesson is for publishers: verify the people pitching you, not just the quality of their prose.

Sources

1. OpenAI, “Disrupting AI-enabled ‘false front’ operations”, 8 October 2026 (company threat-intelligence report; OpenAI’s own findings and ratings). https://openai.com/index/disrupting-ai-enabled-false-front-operations/
2. NPR, “OpenAI bans Russian, Iranian accounts over influence efforts”, 8 October 2026 (news report). https://www.npr.org/2026/10/08/nx-s1-5995576/openai-russia-iran-influence-operations-chatgpt
3. CyberScoop, “OpenAI says Iran, Russia used AI journalists, think tanks to influence Western media”, 8 October 2026 (trade press). https://cyberscoop.com/openai-disrupts-russia-iran-ai-influence-operations/