OpenAI disclosed on Thursday that it identified and suspended accounts operated by Russian and Iranian operatives who were leveraging the company’s AI tools to conduct influence campaigns in various regions. The investigation revealed that these operations employed different tactics to advance their respective agendas and measure their effectiveness.
The Russian operations concentrated on Latin American countries with objectives including undermining support for Ukraine and shaping local political landscapes. Internal communications generated using ChatGPT showed that Russian operatives primarily used the AI system to draft reports documenting their activities for superior officials. One particularly significant campaign involved impersonating a regional educational authority in Peru, resulting in schools organizing events centered on Stepan Bandera, a World War II-era figure whose legacy remains contested across Eastern Europe. OpenAI documented that a Peruvian news outlet reported on one such school event, which subsequently triggered a response from a far-right Polish member of the European Parliament calling for entry bans related to Bandera supporters.
The Iranian-affiliated accounts employed AI differently, utilizing it to generate content in both Persian and English focused on U.S.-Iran relations and American politics. These operatives created fictitious author personas to pitch articles to international media outlets, successfully publishing approximately 100 articles through this method. Several outlets including Daily Kos and the U.K.-based Middle East Monitor published content from these fabricated authors before the deception was uncovered.
OpenAI’s analysis indicated that campaigns emphasizing placement in established news organizations achieved broader audience reach compared to those relying solely on fake social media accounts. The company noted that while both Russian and Iranian operations used AI to generate certain materials, this represented a minority portion of their overall activities. Researchers studying online influence operations characterized the use of AI as enabling operatives to create seemingly authentic content while reducing manual work requirements.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI