Russia is planning a new stage of digital repression: in 2026, Roskomnadzor is to launch an AI-based internet traffic filtering system, which will strengthen censorship and complicate bypassing blockages. This was reported by the Foreign Intelligence Service, writes UNN.
In 2026, Roskomnadzor is considering launching a new mechanism for filtering internet traffic based on machine learning and AI – another step by the Kremlin to deepen digital isolation. USD 29.2 million has already been allocated for the project, despite the chronic budget deficit and the increasing cost of maintaining the existing control infrastructure.
The initiative aims at more stringent and targeted content blocking: from automatic detection of "mirrors" of prohibited resources to neutralizing VPN services and analyzing traffic by content – texts, images, videos, and audio. This involves a transition from crude mass blocking to pinpoint filtering, which significantly complicates bypassing restrictions and intensifies the persecution of internet dissidents and critics of the regime.
Technically, the system will rely on DPI solutions already deployed in the Russian Federation, with the addition of AI modules for real-time analysis of patterns and anomalies, including in conditionally encrypted traffic. The authorities are already using the "Oculus" and "Vepr" systems, which blocked 1.289 million resources in 2025 and restricted access to 5.5 thousand addresses daily. However, their effectiveness is only about 60% and requires increasingly high computational costs.
The new initiative demonstrates the Kremlin's desire not just to strengthen censorship, but to build total control over the information space. At the same time, financial constraints and high technological complexity cast doubt on Moscow's ability to fully implement these plans.