On June 22, the government commission on legislative activity approved a bill to support the development of artificial intelligence in Russia. This was reported by Interfax, citing the press service of Deputy Prime Minister Dmitry Grigorenko.
The bill applies only to large foundational models with over 1 billion parameters. These models are categorized as "sovereign," developed entirely by a Russian legal entity using local infrastructure, and "national," which partially utilize open-source components.
"We plan to implement such solutions in the most sensitive areas—such as government administration. Sovereign and national models will receive priority support from the state," noted Deputy Prime Minister Grigorenko.
According to an unnamed source from Kommersant, the category of "trusted" models for critical information infrastructure was dropped due to existing requirements from the Federal Service for Technical and Export Control (FSTEC) and the Federal Security Service (FSB).
The final version of the bill also removed mandatory labeling of synthesized AI content, sections on copyright, regulations for AI data centers, and restrictions on foreign neural networks. Grigorenko stated that the project does not propose a ban on foreign solutions.
The document is expected to be submitted to the State Duma by the end of the week. Its main provisions are set to take effect on September 1, 2026, while regulations regarding government powers, including the application of models, their definitions, and developers' responsibilities, will come into force on March 1, 2027.
For existing models that do not meet the new criteria, a transition period until September 1, 2032, is provided, as long as the AI data is processed in Russia.
It is worth noting that at the 2025 CNBC East Tech West conference, participants described sovereign AI as the ability of a state to independently control its own technologies, data, and infrastructure.
Earlier, ForkLog reported that not only developing countries but also authorities in the European Union, the United States, and the United Kingdom are turning to the creation of national AI systems and local computing capacities.