The working group of the Civil Code Codification Council under the leadership of the President of the Russian Federation has prepared amendments to the Artificial Intelligence Law and the Civil Code to transform the labeling of generated images, audio and video from a voluntary initiative category to a mandatory requirement.
The initiative proposes lifting previously existing relaxations and allowing users or companies to independently decide whether to issue warnings about the use of neural networks. The procedure, content, and exact format of such marking will be assigned to the authorized government agency, but a specific agency has not been designated in the document. Meanwhile, the current version of the law only imposes technical obligations on large platforms with audiences of more than 500,000 people to provide users with the opportunity to mark appropriately, and these norms are scheduled to come into force on March 1, 2027.
Business representatives and subject matter experts have criticized the new proposals, citing their immaturity and technical impracticality. Market participants said that mandatory labeling will not prevent malicious deepfakes from continuing to use foreign resources, but will instead bring unnecessary obstacles and uncertainty to domestic services and platforms.
