In Russia, a preliminary draft of a national standard for data synthesis is being developed on the basis of the Big Data Association and with the participation of Sber. The goal of the standard is to improve the accessibility, security, and quality of data needed for the development of artificial intelligence.
The main objective of the standard is to describe the technology for creating confidential synthetic data, which will allow the development of AI technologies with privacy compliance at all stages of synthesis.
The document presents mathematical proofs confirming that compliance with the standard's recommendations allows for the synthesis of data without the risk of confidential information leakage. Security is achieved by finding the optimal balance between privacy protection and the quality of the resulting datasets.
Synthetic data is becoming a real alternative to anonymized data, which today is often constrained by excessive regulatory restrictions. When privacy requirements are met, synthetic data does not carry risks and opens a breakthrough path to achieving the goals of data accessibility needed for training artificial intelligence
Experts believe that the approval of the standard in 2025 will be an important step towards integrating synthetic data into wide circulation in the country.
Read more on the topic:
Sber Releases GigaChat Max Neural Network
Yandex invents a way to compress neural networks
Russia Develops Index for Assessing the Ethics of AI Systems in Medicine