Russia Introduces Standard for Synthetic Data Used in AI
The Big Data Association and the National Technology Center for Digital Cryptography have developed a preliminary national standard for synthetic data, artificially generated information that mimics real-world data without revealing individuals' identities.
Russia previously had no unified requirements for synthetic data, despite its use in training AI systems, testing machine learning models, and sharing information between organizations. The document covers neural network-based synthesis methods, privacy protection techniques, and criteria for evaluating data quality.
Yandex, VK, Sber, T-Bank, MTS, and the National Technology Center for Digital Cryptography have worked on the standard since early 2025. The association says such a framework is essential for effectively training Russian AI models: high-quality synthetic data is not linked to the original personal information and can unlock datasets that were previously inaccessible because of their sensitive nature.
Rostelecom, MegaFon, and Sberbank have backed the initiative, saying the standard could reduce the risk of data leaks and accelerate AI adoption across government and business.
The document will remain in effect for three years, after which it could become a full national GOST standard. There is currently no unified international standard in this field, although work is underway at the International Organization for Standardization and the International Electrotechnical Commission.
