OpenAI says Russian and Iranian networks used ChatGPT in influence campaigns

OpenAI says it banned Russian and Iranian accounts that used ChatGPT in influence campaigns across several countries. NPR’s account describes a striking split: Russian operatives mostly used the chatbot to report their work to superiors, while Iranian operatives used it to help produce material for publication.

OpenAI Watch analysis

What happened

NPR reports that OpenAI identified Russian operations targeting countries in Latin America, apparently seeking to undermine support for Ukraine and influence local politics. One operation allegedly impersonated a regional education authority in Peru and persuaded schools to hold activities about Ukraine. OpenAI said the operatives’ own reports described the campaign; NPR also cites a Peruvian outlet reporting that a school event took place.

The Iranian accounts, according to OpenAI’s findings as reported by NPR, used seven fake authors to pitch and polish long articles in Persian and English. They published almost 100 articles, including pieces in Daily Kos’s community section and the Middle East Monitor. OpenAI said the campaign appeared consistent with a commercial influence operation but that it could not identify who was behind it. Iranian accounts also used AI to produce content; in the Russian campaign, OpenAI said that was a minority of the workload.

Why it matters

The report suggests AI’s value in influence campaigns may be less about generating an ocean of fake posts than helping people make deceptive material sound convincing, or appear more polished across languages. NPR quotes Clemson University researcher Darren Linvill saying the technology helped content look authentic enough to be published by legitimate outlets.

There are important limits to the account. These are claims from OpenAI’s investigation, relayed by NPR, not a complete independent accounting of who was responsible for every reported action. NPR also notes that Russian operatives sometimes took credit for events they were not responsible for. The Iranian operation’s sponsor remains unidentified, and Daily Kos said its community posts are separate from staff reporting and are not reviewed before publication.

Our read

This is a substantial account of how AI can slot into an influence operation’s existing machinery, from drafting and translation to reporting back to whoever is paying. The useful distinction is that assistance with persuasive language is not the same as an AI system independently running a campaign. The attribution and the gaps matter here; so does the reported route from a fake byline to a real publication.

What to watch

• Whether OpenAI publishes further evidence or clarifies how it attributed the accounts and activities.
• Whether affected outlets explain how the reported articles passed their publishing checks.
• Whether investigators identify who commissioned the Iranian campaign.

Discussion spark: When a fake author’s AI-assisted article reaches a real publication, should the main responsibility fall on the platform that hosted it, the outlet that published it, or the people running the operation?

OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.