How the Propaganda Machine Operated
OpenAI has revealed that several Russian and Iranian accounts were employing its language models to produce propaganda content under the guise of „fake journalists.” The company banned these accounts after discovering coordinated campaigns aimed at shaping public opinion. The discovery came during routine monitoring of user activity, which flagged suspicious patterns of mass content generation.
What Happens to the Banned Users?
OpenAI’s investigation showed that the accounts were using ChatGPT to write articles, social media posts, and even news briefs that mimicked reputable journalism. The content was tailored to support government narratives, spread misinformation, and influence political discourse in both countries. The company stated that it had no prior knowledge of these activities and that it had acted swiftly to prevent further dissemination.
The banned accounts were part of organized groups that leveraged the AI’s natural‑language capabilities to create large volumes of text quickly. Analysts noted that the content often contained subtle propaganda cues, such as selective framing of events and the use of emotionally charged language. By presenting the material as if it came from independent journalists, the groups sought to legitimize state‑sponsored messages. OpenAI’s internal tools flagged repeated use of specific phrases and the rapid spread of identical content across multiple platforms, which triggered the ban.
Frequently Asked Questions
OpenAI has removed the offending accounts and revoked their access to all of its services. The company also issued a public statement warning other users about the risks of using AI for disinformation. While the bans are permanent, OpenAI has offered to review any new accounts that may have been created by the same individuals. The organization is also working with international partners to identify and mitigate similar threats in other regions.