OpenAI says it banned ChatGPT accounts linked to Russian and Iranian influence campaigns that hid behind a think tank and invented journalists. For readers, the concern is material reaching them through apparently credible publishers with its origins concealed. The company disclosed its findings on October 8, 2026, in Disrupting AI enabled “false front” operations.
The Iranian campaign used seven journalist identities, with nearly 100 articles published or syndicated under their names, according to OpenAI. Its social media replies attracted little engagement. The Russian campaign apparently employed people in Latin America who didn't know who controlled their research organization, the company says.
Examples under the named bylines include Eurasia Review's March 18, 2026, opinion article “You Cannot Bomb Iran Into Regime Change,” credited to Ericka Feusier, and Middle East Monitor's “When peace becomes a media battlefield,” credited to Ervin Hoskins. Haksöz Haber published a Turkish translation of the latter on June 18, explicitly crediting Middle East Monitor as its source, a republication rather than a separate original submission. OpenAI says ChatGPT refined article drafts and wrote editorial pitches; it doesn't quantify the AI contribution to each published piece, and its total combines publication and syndication.
Reach doesn't prove persuasion
OpenAI rates the Russian campaign at Category 5 on the Breakout Scale and the Iranian article campaign at Category 4. Those labels need explaining before anyone treats them as a measure of changed minds.
The framework, published by Brookings in September 2020 as The Breakout Scale: Measuring the impact of influence operations, sorts influence campaigns by observable spread across platforms and communities. Category 4 means material receives mainstream media amplification; Category 5 means prominent people amplify it. The highest category includes a policy response or other concrete action, or a call for violence. It's a way to compare distribution using evidence researchers can check, rather than a count of people persuaded.
There's another reason for caution: OpenAI says the operators inflated their own performance reports. That leaves editors with a practical verification question before publication: who is actually submitting the work? Readers deserve that answer before a convincing biography becomes a substitute for checking a contributor's identity.
The publication pages establish that the bylines reached readers, but they don't explain how editors verified the contributors or whether they knew of any AI assistance. No publisher response or independent assessment of this campaign was located in the searches for this revision. No direct requests for comment were sent.
Sources
- 1.Disrupting AI-enabled “false front” operations · OpenAI
- 2.The Breakout Scale: Measuring the impact of influence operations · Brookings
- 3.You Cannot Bomb Iran Into Regime Change · Eurasia Review
- 4.When peace becomes a media battlefield · Middle East Monitor
- 5.Barış, medyanın savaş alanı haline geldiğinde · Haksöz Haber
Reported by the WattsUpNext desk from the sources linked below. Spot an error? Tell us at corrections@wattsupnext.com.
The WattsUpNext Brief
Get stories like this in your inbox.
One email each weekday, only the topics you choose. Real news, sourced, no fluff. Unsubscribe in one click.
