In Russia, AI Was Made a “Law-Abiding Citizen” in 2025
From a legal standpoint, AI regulation in Russia developed largely as an experimental process.
In 2025, artificial intelligence regulation in Russia continued to evolve without a single overarching “big” law. The legal framework for AI use still rests primarily on the National AI Development Strategy through 2030.
The strategy sets the overall direction for deploying AI technologies in the economy and public administration. More specific rules are shaped through by-laws, standards, and sector-specific regulations.
An Experimental Approach
One of the key legal instruments of the year was the use of experimental legal regimes. These allow companies and government bodies to test AI in real-world conditions while temporarily easing certain regulatory requirements. As a result, AI is already being used in fintech, urban services, and transport. Moscow continues to operate a special regime for AI projects, where technologies are piloted in healthcare, utilities, and city management. Successful solutions are then scaled up.
Moving Toward Standards
Another notable trend was the standardization of data practices and AI systems themselves. In 2025, Rosstandart expanded its set of GOST standards for artificial intelligence. These cover requirements for model quality, lifecycle management, terminology, and applied use cases. The standards are already being used in public-sector procurement and major infrastructure projects, making it easier to accept new solutions and reducing implementation risks.
Trust and Security
A third area focused on trust and security. AI is increasingly treated as a standard part of IT infrastructure and subject to the same approaches as other critical systems. These include access control, activity logging, testing, and incident readiness. Banks, telecom operators, and public authorities are already applying these practices when launching AI services.
The overall picture for 2025 looks like this: rather than rushing into strict regulation, the state is building rules through pilot projects, industry standards, and real-world practice. This approach allows AI to develop without sudden regulatory shocks, while at the same time reducing risks for businesses and everyday users.
