MOSCOW, RUSSIA / RankWire.AI / – The Federation Council sanctioned a comprehensive bill establishing national standards for artificial intelligence on July 17, outlining regulations for large foundation models within Russia. The legislation specifies the technologies it covers and grants authority to government agencies. It also sets standards regarding model ownership, local data storage, transparency for users, and AI-generated content. Having passed the State Duma on July 8, the bill now awaits presidential approval and official publication before becoming law at the federal level.
The bill describes a large foundation model as software capable of executing multiple intellectual tasks at a level comparable to humans. To qualify, a system must include no less than 1 billion parameters. Such systems can provide information, make decisions, or predict outcomes based on human-set goals. The framework emphasizes principles around technological sovereignty, human rights, personal choice, security, and adherence to Russian legislation. These principles govern the development, deployment, and use of qualifying AI technologies.
The legislation introduces categories of sovereign and national models linked to Russian control. A sovereign model must be developed by a Russian legal entity and operate within data centers located inside Russia. Its creators must be able to reproduce the entire development process, including training and original parameters. A national model adheres to similar ownership and localization criteria but can incorporate foreign software components released under open licenses, provided Russian entities maintain control and operational authority.
Legal Designations for Domestic AI Systems
The government may support developers involved in creating, deploying, or managing qualifying foundation models. This support could include access to state-owned datasets for training purposes. Authorities might also mandate exclusive use of sovereign or national models within government information systems and other sensitive environments. Additional rules concerning defense, security, public order, and property protection might be established through separate laws or presidential decrees. These regulations are assigned to specific government bodies to enforce within their legal jurisdictions.
Digital service providers with more than 500,000 daily users face a separate obligation concerning AI-generated audio and visual content. Such platforms must offer a feature enabling users to label or mark this type of material. This applies to websites, apps, and social media platforms. The regulation does not require all content to be automatically labeled by the platform; instead, developers and users can agree on the labeling method through service agreements. The key objective is to provide an option for users to disclose or identify qualifying AI-generated content.
Standards for Copyright and Content Disclosure
AI service providers are mandated to inform users about the ownership rights of generated content. They must also clarify access conditions and whether the content may be downloaded or transferred. The legislation separately addresses the use of copyrighted works for machine learning. It permits analysis for extraction, comparison, classification, and pattern recognition when developers have lawful access. Training with protected works is allowed if no technical restrictions are bypassed. The rules connect model training practices to existing copyright and access regulations.
The majority of these regulations are set to become effective on September 1, 2026, after presidential approval and official publication. Specific rules regarding domestic model classification, developer responsibilities, content marking, and intellectual property will take effect from March 1, 2027. Existing systems may operate until September 1, 2032, provided they process and store data within Russia. Until the legislation is officially signed and published, it remains an approved bill rather than an enacted law in Russia’s legislative process.