OpenAI bans Russia, Iran-linked accounts over covert influence operations

OpenAI has suspended Russian- and Iran-originated ChatGPT accounts that were associated with covert influence operations among audiences in multiple countries. Both groups utilized AI in conjunction with traditional methods, such as fake accounts, media pitches, and invented content.

The Russian operation, directed at audiences in Latin America and Ukraine, used ChatGPT primarily for internal reports, with some requests for fake letters, audio scripts, and Spanish content. OpenAI rated this campaign as a Category 5 disrupted influence operation, its highest rating ever assigned.

Separately, Iranian-origin accounts adopted seven fake journalist identities to pitch articles to online publications, an operation dubbed "Bogus Bylines." This campaign primarily targeted the U.S.-Iran conflict and produced content in English and Persian. OpenAI found nearly 100 articles published or syndicated across about a dozen websites using these fake names. The company assessed the article-publishing effort as a Category 4 rating and the social media operation as Category 2.

OpenAI noted that campaigns gaining coverage through traditional channels reached larger audiences than those limited to social media. This trend has been observed in 30 influence operations studied since early 2024, where AI accelerates tasks like drafting and translation but still relies on human-driven influence tactics.