MOSCOW, RUSSIA / RankWire.AI / – Russia’s Federation Council approved a legislative bill on July 17 that establishes national guidelines for the development and deployment of large artificial intelligence foundation models. The legislation defines key legal terms, grants regulatory authority, and offers support for domestic developers that meet qualification criteria. It also outlines requirements related to model governance, data storage, user notifications, and AI-generated content. The bill successfully passed the State Duma on July 8 and is now pending presidential approval before becoming law.
According to the bill, a large foundation model is software capable of performing numerous intellectual tasks at a level comparable to human performance. The definition encompasses systems that deliver information, make decisions, or predict outcomes based on human-set objectives. A qualifying model must have a minimum of 1 billion parameters. The framework emphasizes principles such as technological sovereignty, human rights, individual choice, security, and legal compliance within Russia.
The legislation distinguishes between sovereign and national model categories. A sovereign model must be developed by a Russian legal entity and remain under local control, utilizing data centers located within Russia. Developers are required to retain the technical ability to reproduce the entire development process, including training and the original parameters. Meanwhile, a national model follows similar ownership and data handling rules but can incorporate foreign components licensed under open-source agreements.
Domestic models to receive official recognition
The government is permitted to implement support measures for companies involved in creating, deploying, or managing qualifying foundation models. Such measures may include access to state-held data necessary for training. Additionally, authorities can designate specific environments where only sovereign or national models are permitted to operate, particularly for government information systems and other sensitive sectors. Separate legislation and presidential decrees may establish rules concerning defense, security, public order, and property protection.
Websites, applications, and social media platforms will face a new content regulation. Platforms with more than 500,000 daily active users are required to provide a mechanism for marking AI-generated audio or visual content. Users will be able to add an information notice to qualifying material. The format of these labels will be determined by developers and users through service agreements. While automatic labeling of every item is not mandatory, platforms must offer the technical capability to do so.
Intellectual property and enforcement provisions outlined
AI service providers are required to inform users about ownership rights related to generated content. They must also clarify access conditions and whether users can transfer or reproduce the material. The law also addresses the use of copyrighted works for machine learning, stating that analysis activities such as extraction, comparison, classification, and pattern detection are lawful when developers have legally obtained access. Training on protected works is permitted when no technical restrictions were bypassed to access the content.
Most provisions are set to take effect on September 1, 2026, contingent on presidential approval and publication. Specific rules concerning domestic model status, developer responsibilities, content labeling, and intellectual property rights will become effective on March 1, 2027. Existing AI systems will benefit from a transition period until September 1, 2032, during which their data processing and storage must remain within Russia. Until the measure is officially enacted and published, it remains an approved bill, not a law.