Russia AI Regulation Overview

Overview

Russia's approach to artificial intelligence regulation combines centralized, state-directed strategic planning with flexible experimental sandboxes and voluntary ethical principles. The overarching regulatory vision is established in the National Strategy for the Development of Artificial Intelligence until 2030, originally approved by Presidential Decree No. 490 in October 2019 and substantially updated by Presidential Decree No. 124 in February 2024. The strategy emphasizes technological sovereignty, import substitution in hardware and software, expanding compute capacity, and achieving global leadership in targeted AI applications.

To transition AI technologies from research to commercial adoption while managing risks, Russia relies on regulatory sandboxes known as experimental legal regimes (ELRs). Framework Law No. 258-FZ (2020) and its amending legislation, Federal Law No. 169-FZ (2024), permit temporary exemptions from existing regulatory restrictions for qualified participants. This enables controlled testing in real-world conditions across priority sectors like healthcare, autonomous transportation, precision agriculture, and financial technology under mandatory government oversight and compulsory civil liability insurance.

Regulatory Approach

Russia employs a hybrid regulatory approach that pairs targeted, risk-oriented legal mechanisms for experimental deployment with soft-law ethical self-regulation. Rather than imposing comprehensive, rigid statutory rules across all AI deployments, the state creates controlled environments via experimental legal regimes where specific statutory requirements can be temporarily modified or suspended. This sandbox model allows regulators to collect empirical data on AI safety and efficacy before adapting permanent national legislation.

Complementing this sandbox mechanism is a voluntary soft-law framework embodied in the Code of Ethics in the Field of Artificial Intelligence (2021). The Code sets high-level ethical principles—such as human-centredness, non-discrimination, precaution, transparency, and risk mitigation—to guide developers, public bodies, and users in non-military domains. Safety and liability are reinforced through mandatory civil liability insurance for sandbox participants operating AI tools, along with administrative procedures for investigating AI-related incidents.

Key AI Legislation

• National Strategy for the Development of Artificial Intelligence until 2030 (Presidential Decree No. 490 of 2019, updated by Presidential Decree No. 124 of 2024): Serves as the primary national blueprint defining strategic goals, technical definitions (including foundation and generative models), technological sovereignty mandates, compute targets, and sector readiness goals through 2030.
• Federal Law No. 258-FZ on Experimental Legal Regimes in the Sphere of Digital Innovations (2020): Establishes the legal sandbox framework allowing legal entities and individual entrepreneurs to test AI and digital innovations under temporary regulatory exemptions for up to three years.
• Federal Law No. 169-FZ (2024): Amends Law No. 258-FZ to establish mandatory civil liability insurance for sandbox participants using AI technologies, streamline the ELR proposal review process from 90 to 50 working days, introduce incident investigation mechanisms, and clarify intellectual property usage rules.
• Code of Ethics in the Field of Artificial Intelligence (2021): A national voluntary ethical framework setting principles for human dignity, fairness, risk orientation, transparency, and organizational accountability across the AI lifecycle.

Governance & Enforcement Bodies

Governance of AI across Russia involves federal executive authorities, inter-agency councils, special competence hubs, and state corporations. The Russian Government holds primary responsibility for strategy execution. The Ministry of Economic Development functions as the authorizing federal body administering the experimental legal regime framework, maintaining the Special Registry of sandbox participants, and coordinating cross-sectoral AI initiatives with sectoral regulators.

Other key bodies include the Ministry of Digital Development (managing digital infrastructure, telecommunications, and compute allocation), the Ministry of Education and Science (coordinating university AI research and education), the Analytical Center for the Government, and the Federal AI Center. The inter-agency Coordination Council reviews sandbox applications, approves regulatory exemptions, and monitors regime outcomes. Additionally, major state corporations (such as Sberbank, Rostec, and Rosatom) lead sector-specific AI integration efforts under national policy targets.

Penalties & Enforcement

Enforcement within the legal corpus operates primarily through administrative oversight, registry management, and participant status controls within experimental legal regimes. Under Federal Law No. 258-FZ and Federal Law No. 169-FZ, participants that fail to comply with regime programs or cause harm during testing may face suspension of status or revocation of their entry in the Special Registry maintained by the Ministry of Economic Development. A dedicated commission is empowered to investigate incidents where AI operating within an ELR causes harm to life, health, or property, requesting documentation to determine preventative or corrective measures.

To guarantee financial redress, Federal Law No. 169-FZ mandates that participants in AI-focused sandboxes maintain civil liability insurance covering potential third-party damage to life, health, or property. Participants must also demonstrate clean criminal records regarding economic crimes (for specified activity areas) and share joint safety responsibility with supervising bodies. The supplied corpus contains no statutory tables of administrative monetary fines or criminal penalty codes dedicated specifically to general AI non-compliance outside of these sandbox administrative and civil liability controls.

Data Protection Framework

Data protection principles for AI systems are outlined in the National AI Strategy, the Code of Ethics, and sandbox implementation rules. The National AI Strategy mandates the creation of national training datasets across priority domains while explicitly enforcing personal data protection and privacy safeguards. Under the Code of Ethics, organizations processing data for AI systems are expected to maintain lawful processing practices, enforce strict data security measures, prevent algorithmic bias, and ensure transparency when interacting with individuals.

Within experimental legal regimes, Law No. 169-FZ requires sandbox subjects to maintain a registry of individuals who enter into legal relations with them involving solutions developed using AI technologies. Furthermore, sector-specific deployment rules—such as video analytics in smart city public safety or medical data usage in diagnostic platforms—require balancing algorithmic processing with privacy protection. The provided corpus does not include details on a general omnibus data protection statute, focusing instead on data governance provisions embedded directly within AI regulation records.

Sector-Specific Rules

Sector-specific AI deployment is executed primarily through targeted roadmaps under the National AI Strategy and approved sandbox programs under Federal Law No. 258-FZ. In healthcare, rules facilitate testing AI diagnostic systems analyzing medical imaging and laboratory data, preventive examination tools integrated into the Unified State Health Information System, and biobank analytics platforms. In transport and logistics, frameworks support autonomous vehicle testing on roads, last-mile logistics robots, smart city traffic management, and predictive maintenance for railway infrastructure.

Other priority sectors with tailored implementation provisions include financial technology (AI credit scoring, fraud detection, RegTech compliance, automated investment advisory), precision agriculture (satellite imagery analysis for crop yields, livestock monitoring, autonomous tractors), smart manufacturing (predictive maintenance, computer vision quality control, cobots), and municipal smart cities (energy grid optimization and public safety analytics). Key state entities lead sectoral integration, such as Sberbank for financial AI, Rostec for industrial and defense applications, and Russian Railways for transportation systems.

International Alignment

Russia's international strategy for AI focuses on participating in global governance initiatives, engaging in international technical standardization, and expanding technological cooperation with friendly nations while emphasizing domestic technological sovereignty. The National Strategy prioritizes advancing Russian AI solutions in international markets and mitigating the impact of Western sanctions by establishing independent domestic hardware, software, and compute capabilities.

While the Code of Ethics in AI (2021) aligns with international high-level principles such as human-centredness, precaution, non-discrimination, and transparency, the provided corpus contains no explicit reference to direct adoption of or alignment with specific foreign statutory frameworks like the EU AI Act or OECD AI Principles. Strategic focus remains centered on technological self-reliance, import substitution, and bilateral partnerships that safeguard strategic capabilities.

Future Developments

Future regulatory and implementation developments focus on fulfilling the mandates set out in the Updated National AI Strategy (Decree No. 124) through 2030. Key upcoming milestones include the mandatory integration of the federal AI project into the national Data Economy project by July 1, 2024, and a formal mid-term strategic evaluation scheduled for December 31, 2027. Federal executive bodies and sectoral regulators are instructed to continually submit normative proposals to eliminate administrative barriers and operationalize strategy goals.

On the legislative front, Federal Law No. 169-FZ entered into force on January 4, 2025, enforcing mandatory civil liability insurance for AI experimental regimes and activating shortened 50-day approval timelines for sandbox proposals. Quantitatively, the strategy targets training 6,000 AI specialists annually by 2030, increasing top-tier research publications to 500 per year, expanding domestic high-performance compute capacity, and broadly deploying foundation and generative AI models across key economic sectors.

Enforcement Bodies