OpenAI shuts Russian and Iranian accounts that used ChatGPT for propaganda and journalist impersonation
OpenAI blocked accounts linked to Russian and Iranian influence operations that used ChatGPT to draft geopolitical propaganda, impersonate journalists, and assist covert organizations, according to NBC News.
OpenAI has closed accounts tied to Russian and Iranian influence campaigns that were using its ChatGPT system to generate political content and impersonate journalists, according to reporting by NBC News.
These accounts were part of organized efforts to use ChatGPT for drafting propaganda narratives, adjusting messages for different audiences, and preparing content under false or stolen identities. According to the reporting, one operation linked to Iran created seven fake personas presented as Western journalists to place articles and commentary, while Russian‑aligned accounts used the tool to support messaging favorable to Moscow’s positions.
The removals illustrate how generative AI platforms have become tools in information campaigns. Systems like ChatGPT can turn prompts into coherent text on almost any topic in many languages. For influence operators, that capability helps replace or augment teams of human writers, enabling them to respond quickly to events and maintain a steady flow of tailored material.
When those campaigns involve fabricated reporter identities, the damage goes beyond individual posts. Fake journalists exploit the trust normally associated with bylines and professional credentials, and they make it harder for editors and readers to distinguish genuine reporting from state‑aligned influence content.
Once AI‑assisted drafts leave the platform, they can appear in many formats: social media posts, articles on marginal outlets, or commentary attributed to supposed analysts. By the time such content spreads, the fact that it originated from prompts fed to an AI model by a state‑linked operator is no longer visible.
OpenAI’s action reflects a broader shift in which technology firms are being drawn into a gatekeeping role against state‑backed disinformation. Detecting and disrupting misuse by foreign actors requires internal monitoring and coordination with outside stakeholders, which in turn raises questions about how these systems are overseen and how decisions are communicated.
The episode also shows how influence tactics are evolving. Instead of relying only on networks of human trolls or simple automated bots, Russian‑ and Iranian‑linked operations are now folding generative AI into their workflows and then pushing the resulting material across different platforms and outlets.
Developments to watch include further disclosures from OpenAI or other AI providers about comparable influence efforts, any public guidance from governments on how AI platforms should respond to state‑linked misuse, and changes in platform policies on political content. If other major AI services report similar account takedowns, or if observable propaganda networks shift to new tools, that will indicate how both operators and platforms are adapting.
Sources
- OSINT