Major AI chatbots directed a significant volume of traffic to Russian state-affiliated and pro-Kremlin websites during the final months of 2025, according to independent web analytics data.
Between October and December, more than 300,000 visits were traced back to AI-generated referrals pointing users toward eight propaganda-linked domains. These visits came from widely used tools such as ChatGPT, Perplexity, Claude and Mistral.
Key findings from the data include:
- RT-related domains received the highest share of AI-driven referrals
- Large Russian outlets like RIA, Lenta, Izvestia and Rossiyskaya Gazeta followed
- Smaller propaganda sites saw fewer visits but higher proportional impact
On several niche sites, AI chatbot referrals made up between 6% and nearly 10% of total referral traffic. For major state media platforms, however, AItraffic represented only a small slice of their overall audience.
Researchers also highlighted a Moscow-based disinformation operation known as Pravda, which publishes massive volumes of pro-Kremlin content aimed at influencing AI systems. The network reportedly released millions of articles in a single year to overwhelm search indexes and AI retrieval models.
Audits showed that leading AI tools repeated false narratives tied to this network in a notable portion of tested responses, often without clear labels or warnings. Public statements from individuals linked to the operation suggest the goal is to shape global information flows by embedding these narratives directly into AI-generated answers.
