OpenAI Bans Russian and Iranian Accounts That Used ChatGPT to Plant False Stories in News Outlets

OpenAI said in a blog post on Thursday that it has banned Russian and Iranian accounts that were using its artificial intelligence tools to help run influence operations in multiple countries. The operatives claimed in internal reports, written with ChatGPT itself, that they had managed to get false narratives picked up by news outlets.

OpenAI investigated the impact of the campaigns the accounts described. It found that those aimed at placing false narratives in established media outlets appeared to have reached a wider audience than campaigns built on fake social media accounts.

The Russian operations targeted countries in Latin America and appeared aimed at undermining support for Ukraine and influencing local politics, the company said. The accounts mostly used ChatGPT to write reports to unnamed superiors about their work. OpenAI said one of those reports described what it called the widest-reaching influence operation it has come across, because it produced a public response from a political figure.

In that campaign, the operatives said they impersonated a regional educational authority in Peru and tricked some schools into organizing activities about Ukraine that featured Stepan Bandera. Bandera is a controversial figure, admired by nationalists for leading a drive for Ukrainian independence before and during World War II. He is reviled by others, including in Poland, because his movement sided with the Nazis and killed Poles and Jews. The Peruvian outlet Extra reported that one such event did take place at a school.

A Polish news report about the school event then prompted a far-right Polish member of the European Parliament to suggest banning the entry of people who profess support for Bandera. "The Russian fake thus both fed on, and fed into, historical tensions between Ukrainians and Poles," OpenAI wrote. The Russian memos also claimed the operatives ran a fake think tank in Latin America and recruited unwitting researchers to interview experts and compose reports. OpenAI compared that to a 2020 operation posing as an independent news outlet that was linked to the late oligarch Yevgeny Prigozhin's Internet Research Agency.

The Iranian accounts used AI to push out content in Persian and English, mainly about the U.S.-Iran war and, to a lesser degree, American politics. OpenAI said the activity looked consistent with a commercial actor running a for-hire influence campaign, but it could not identify who that was. The accounts invented seven fake authors to pitch and polish long articles to small outlets around the world, publishing almost 100 pieces in this way.

Those outlets included the progressive U.S. site Daily Kos and the U.K.-based Middle East Monitor. According to NBC News, the Middle East Monitor published at least 21 articles by four fake authors. The outlet did not respond to a request for comment and has since taken down the author pages. Daily Kos told NPR that its community section, where one of the personas posted, is separate from staff reporting and that posts are not reviewed by staff before publication. The persona's posts are still up.

Darren Linvill, co-director of Clemson University's Media Forensics Hub, said Russia is constantly engaged in such "active measures." He noted the Russian operatives used AI as a shortcut, "in the same way that a lot of us use AI, but in this case it helped get them caught." Of the Iranian effort, he said AI helped produce language that read as authentic enough to be published by legitimate outlets, something an Iranian operation "would have struggled a lot more with" a few years ago.