MOSCOW, RUSSIA / RankWire.AI / – The Federation Council ratified a legislative framework for artificial intelligence on July 17, establishing guidelines for large foundation models within Russia. The legislation outlines the scope of covered technologies and delegates authority to government agencies. It also sets standards concerning model ownership, domestic data storage, user notification, and AI-produced content. Having been approved by the State Duma on July 8, the bill now awaits presidential approval and official publication before becoming law.
The draft defines a large foundation model as software capable of executing numerous intellectual tasks at a level comparable to humans. A system qualifies if it contains at least 1 billion parameters. Such models can generate information, support decision-making, or predict outcomes based on human-defined objectives. The framework emphasizes principles of technological sovereignty, human rights, individual choice, security, and legal compliance, which apply throughout the development, deployment, and utilization of qualifying AI systems.
The legislation categorizes models as sovereign or national, both linked to Russian control. A sovereign model must originate from a Russian legal entity and operate using data centers located within Russia. Its developers need to maintain the capacity to reproduce the entire development process, including training and the original parameters. A national model adheres to similar ownership and localization requirements but allows the use of foreign software components under open licenses, provided Russian entities retain control and operational authority.
Legal Designations for Domestic AI Models
The government might extend support to developers engaged in creating, deploying, or managing qualifying foundation models. Such support could involve access to state-held 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 related to defense, security, public order, and property protection may be established through separate legislation or presidential decrees. Overall, state agencies are tasked with enforcing these requirements within their jurisdiction.
Large-scale digital platforms face specific obligations regarding AI-generated audiovisual content. Platforms with over 500,000 daily users are required to provide a tool enabling users to mark such materials. This regulation applies to websites, applications, and social media platforms. It does not demand automatic labeling of every piece of content; instead, developers and users can agree on the format of notices through service agreements. The focus is on offering an option for creators and distributors of qualifying content to disclose the nature of their material.
Standards for Copyright and Content Disclosure
AI service providers are required to inform users about the ownership rights of generated content. They must also clarify access conditions and whether users are permitted to download or transfer the material. The bill also addresses the use of copyrighted works for machine learning, allowing analysis for extraction, comparison, classification, and pattern recognition when lawful access has been obtained. Training on protected works is permitted if no technical restrictions are bypassed. These rules connect model training practices with existing copyright and access regulations.
Most provisions are set to become effective on September 1, 2026, following presidential approval and official publication. Regulations concerning domestic model status, developer responsibilities, content labeling, and intellectual property rights will take effect on 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 federal law under Russia’s legislative process.