
Automated Persuasion: How AI Accounts Bureaucratized Foreign Influence Operations
OpenAI shuttered Russian and Iranian networks using ChatGPT to publish syndicated articles and staff covert think tanks.

The expectation that artificial intelligence would primarily serve as a tool for mundane corporate efficiency has once again met the predictable realities of international statecraft. According to findings published by OpenAI, state-linked actors operating from Russia and Iran systematically deployed ChatGPT to establish counterfeit journalistic identities, manage deceptive front organizations, and churn out political commentary.
Among the disrupted networks was a Russian operation that constructed a Latin American think tank titled the Social Research Center. According to OpenAI, the group recruited local personnel who were unaware of their true employer to produce original material aimed at reshaping regional perceptions of Ukraine. OpenAI described this setup as the most complex front identity it had dismantled in two and a half years. Beyond Latin America, the network seized upon European political friction, generating targeted social media reactions around a proposal by Polish far-right MEP Grzegorz Braun to penalize "anti-Polishness," aiming to inflame domestic debates over Poland's unwavering alignment with Kyiv.
Simultaneously, an Iranian campaign relied on seven artificial writer profiles masquerading as Western journalists, including a persona branded as American freelancer Ervin B. Hoskins. OpenAI reported that this network successfully placed nearly 100 long-form articles across roughly a dozen small and medium-sized international affairs outlets between July 2025 and October 2026. Network activity accelerated following the outbreak of the US-Iran war, leveraging automated text generation to distribute commentary across platforms such as X and Instagram while seeking to influence public discourse in Washington and Israel.
These account closures illustrate a clear shift in how digital tools are deployed during geopolitical friction. Rather than replacing traditional propaganda methods, generative models are being integrated into routine operational workflows—writing internal administrative reports, drafting syndicated commentaries, and maintaining fictitious personas across global platforms.
Written by Martina Kirchner martina.kirchner@alpineweekly.com




