OpenAI bans Russian and Iranian accounts that used ChatGPT for influence campaigns, including seven fake authors behind almost 100 articles

OpenAI said on October 8 that it had banned ChatGPT accounts tied to a Russian and an Iranian covert influence operation, NPR reported. Anadolu Agency reported that OpenAI shared what it found with the relevant authorities. OpenAI's own report page refused access from here twice (403), so the details are as the two outlets report them.

The Iranian accounts invented seven fake authors to pitch and polish long articles for small media outlets. NPR reported that they published almost 100 articles this way; Anadolu also reports almost 100, across more than a dozen outlets. According to NPR, OpenAI said it looked like a commercial actor running a for-hire campaign, but it could not identify who.

NPR reported that the Russian operation aimed at Latin America, working to undermine support for Ukraine. Its operators said they posed as a regional education authority in Peru and got schools to run activities about Ukraine that featured Stepan Bandera. Anadolu reports that the operation ran a research platform called the Social Research Center through an invented persona, “Mia Clark”, and that OpenAI said the platform's local staff did not know they were working for a Russian group. NPR reported that the operatives mostly used ChatGPT to write reports for their bosses.

• Reported OpenAI banned Russian and Iranian accounts used for influence operations, NPR reported. NPR via WUNC
• Reported The Iranian operation's seven fake authors published almost 100 articles, NPR and Anadolu Agency report. Anadolu Agency
• Reported The Russian operation posed as a regional education authority in Peru, NPR reported. NPR via GPB