The Russian government is betting on technological independence in the field of artificial intelligence. Prime Minister Mikhail Mishustin has approved a list of instructions following a strategic session dedicated to scientific and technological development. The key vector is consolidating all research support measures into a single management framework.

Tax benefits — into a unified system

The Ministry of Education and Science, the Ministry of Finance, the Ministry of Economic Development, and the Federal Tax Service have been tasked with ensuring a phased transition to providing tax preferences for scientific developments through a unified platform. The deadline is November 27. This is not about introducing new benefits, but about changing the channel for obtaining them: instead of fragmented procedures, agencies will operate in a single digital environment.

In parallel, the system is being filled with industry data. The main budget fund managers financing medical developments must submit information about applied research for 2026. Medicine was chosen first for a reason: the industry already generates vast amounts of data that can be used to train models.

Digital medicine and the legal framework

The MosMedAI platform unites more than two thousand clinics from 75 regions and has processed over 16 million radiology studies. Meanwhile, Yandex has registered a separate legal entity, LLC Yandex Med, for medical software. The legal framework requires fine-tuning: the Ministry of Education and Science, together with interested bodies, will prepare proposals for legislative changes by October 12.

The amendments concern monitoring, quality control, and traceability of information. Transparency here is fundamental — without it, it is impossible to audit expenditures and correctly use data.

Research data — a resource for domestic AI

The second instruction is addressed to the Ministry of Education and Science, the Ministry of Economic Development, Rospatent, and the Ministry of Digital Development. By November 25, the agencies will work out a mechanism for training sovereign and national AI models on data from the unified system. The main condition is reliable protection of intellectual property.

The terms "sovereign" and "national" model are taken from the framework bill on AI regulation, approved by the State Duma in the first reading on July 7, 2026. A sovereign model requires full control and placement in domestic data centers without foreign components, while a national model allows open licenses.

Copyright protection in training neural networks

Data on applications, developments, and research results contains information whose rights belong to specific organizations and authors. The new mechanism will differentiate the availability of datasets for training models so that the process does not devalue patent rights.

The dispute over rights to training data has already unfolded around the framework law. The Russian Book Union, the National Federation of Music Industry, and the Animation Film Association sent letters to authorities demanding the refinement of the provision under which processing protected works for training is not considered a violation if a legal copy is available. Rights holders insist on transparent licensing terms. For scientific data, the issue is more acute: the commercial value of a patentable development directly depends on the regime of its disclosure.

My view: Creating a unified system for managing scientific data is a logical step to accelerate AI development, but the key challenge will remain the balance between data openness and protection of trade secrets. Without clear licensing and traceability rules, this ambitious project risks getting bogged down in bureaucratic disputes rather than technological innovation. However, if the mechanism works, Russia will gain a serious advantage in creating models trained on unique national data.