Roskomnadzor Reports Launch of AI Internet Traffic Filter: System Cost 2.27 Billion Rubles

Roskomnadzor has officially confirmed the deployment of an internet traffic filtering mechanism utilizing machine learning algorithms. This was reported by Forbes, citing the passport of the departmental digital transformation program for 2026 and the 2027–2028 planning period. According to the publication, 2.27 billion rubles were allocated for the development and deployment of the system. The document states that this is a Russian model operating on the agency's own infrastructure; other technical details are not publicly disclosed.

How the System is Structured and Where It Is Integrated

According to the program passport text, the artificial intelligence-based technology is integrated into the automated system for ensuring the security of the Russian segment of the internet (ASBI). The new AI module doesvas not replace existing mechanisms, but complements the operation of technical means of countering threats (TSPU), which already perform network traffic filtering at the telecom operator level. Thus, the agency is establishing a multi-level control architecture: TSPUs operate at the perimeter, while machine learning algorithms add semantic and contextual analysis of the traffic flow.

From Plan to Implementation: Project Chronology

As early as the beginning of 2026, Forbes wrote about RKN's plans to create and implement this exact AI-driven filtering mechanism. At that stage, it was assumed that Roskomnadzor would utilize the Astra AI platform, developed by Astra Group. By November 2026, the project moved from the planning stage to reporting: the agency recorded the fact of implementation and specified the exact amount of expenses. The transition from announcement to confirmed launch took approximately nine months, according to available data.

Budgetary Context: 60 Billion for Three Years

The implementation of the AI filter coincides with a massive increase in budget allocations for internet traffic control. The Russian budget for the next three years has earmarked a record 60 billion rubles for traffic filtering. These expenses are associated with mounting resource blockages and intensified efforts to combat the use of VPN services. The figure of 60 billion for the three-year period makes the filtering system one of the most expensive elements of the state's digital infrastructure.

Previous Initiatives and Expert Assessments

The use of AI in Roskomnadzor's work is not a new direction. Kommersant reported in April 2024 that RKN intended to use artificial intelligence systems to maintain the registry of prohibited information. In October 2026, the Ministry of Digital Development, in turn, proposed using AI to identify and block VPN connections. Experts commenting on these initiatives warned that the use of machine learning models, especially at the initial stage of their operation, carries the risk of mass erroneous blockages of legitimate traffic. These remarks remain relevant regarding the newly launched mechanism as well, since open information on the model's accuracy and mechanisms for appealing false positives is not contained in the program passport.