In addition, the law introduces categories of artificial… (закон · модель · маркировка)
In addition, the law introduces categories of artificial intelligence models such as sovereign and national ones, for which requirements are set regarding the localization of data processing on Russian territory with the involvement of Russian legal entities, as well as support measures for developers of such AI models, including in terms of access to data necessary for training these models,
Consensus
- The State Duma adopted a law on July 8 regulating the use of artificial intelligence in media content creation.
- The law requires large internet platforms with daily audiences exceeding 500,000 to provide technical functionality for marking AI-generated audio and video content.
- Users will be allowed to independently add warnings about the use of neural networks in their content.
- The law introduces categories of 'sovereign' and 'national' large foundational AI models with specific requirements regarding data localization and development support.
- The law applies to content published in Russian or languages of the peoples of Russia.
Points of divergence
- Authors can determine the format and procedure for labeling through agreement with the model developer, and the law includes measures supporting developers of sovereign and national AI models, including access to data needed for training. — vm
- The law is described as framework legislation, allowing future regulation of all AI-related issues via subordinate acts. It explicitly states that using information for training sovereign and national models does not constitute copyright infringement if the material was obtained legally or was in public access. — kommersant
Coverage (2 sources)
- State Duma has adopted a law on mandatory labeling of AI-generated content — Вечерняя Москва
- State Duma has adopted a law on the labeling of AI-generated content — Коммерсантъ