šŸ“œ Media Market Demands Revisions to AI Draft Bill

Russia’s largest media market associations—the Russian Book Union, the National Federation of the Music Industry (NFMI), and the Animated Film Association—have issued a second appeal to the country's top leadership, demanding comprehensive revisions to a draft bill regulating artificial intelligence.

The letters were sent to Anton Vaino, Chief of Staff of the Presidential Executive Office; Prime Minister Mikhail Mishustin; Vyacheslav Volodin, Speaker of the State Duma; and Sergey Boyarsky, Chairman of the relevant parliamentary committee. The authors point out that the feedback they submitted in April was completely ignored in the current draft of the document.

The creative industries’ primary grievance concerns the use of copyrighted material to train neural networks. The current version of the bill effectively exempts the extraction of data from protected works from being classified as an infringement, provided the materials were accessible online for analysis. The authors of the appeal warn that, in its current form, the law will “disincentivize the creativity of domestic authors and turn Russia into a free-of-charge hub for training AI models.” The music sector is particularly alarmed: AI models owned by major IT corporations are training on artists’ lyrics, music, and voices for free, subsequently replicating them without any form of compensation—a practice that directly contradicts the Civil Code of the Russian Federation.

The drafters of the bill partially revised the document prior to its submission by introducing an opt-out mechanism. Under this provision, copyright holders can explicitly ban the use of their content by AI models and file claims if this restriction is violated. However, media market representatives view this measure as insufficient and are demanding the establishment of robust, legally binding mechanisms for the commercial licensing of datasets.

The draft bill itself, which sets out a framework for AI regulation, narrows state support exclusively to large foundational models (exceeding 1 billion parameters), classifying them into “sovereign” and “national” categories. At the same time, the authors have deferred the most contentious issues—including copyright and liability for AI-generated outputs—to future secondary legislation.

The office of Deputy Prime Minister Dmitry Grigorenko confirmed that while the document currently establishes only baseline protection principles, copyright protection norms will be refined through continued dialogue with the industry ahead of the bill’s second reading.

Source: Kommersant