The national standard GOST R 72515, “Transparency Taxonomy of Artificial Intelligence Systems,” has been published on the Rosstandart website. The document was prepared by Volgograd State University and the Institute of the Information Society based on their own Russian translation of the international standard ISO/IEC 12792:2025, “Information technology — Artificial intelligence (AI) — Transparency taxonomy of AI systems.” The standard was submitted by Technical Committee for Standardization TC 164 “Artificial Intelligence” and comes into effect on May 1, 2026.
The objectives of this standard are:
- to enhance trust, accountability, and communication among various artificial intelligence (AI) stakeholders, including supply chain partners, customers, users, society, and regulators, by creating a consistent terminology related to the transparency of AI systems
- to provide AI stakeholders with information about various transparency elements, outlining their significance and potential limitations for different use cases and target audiences
- to provide a basis for developing AI system transparency standards for specific technologies, industries, and geographies
Transparency of AI systems is the ability of stakeholders to obtain adequate information about the system. This information may include information about the system’s functionality, limitations, data, system design, and design decisions. Increased transparency provides relevant stakeholders with information that allows them to better understand how AI systems are developed and used. For example, it allows an AI client (e.g., an AI user) to determine whether the system is suitable for their situation, and it helps an AI auditor assess whether it meets compliance requirements. A standardized transparency taxonomy allows representatives of different professions to better understand each other through the use of common terminology. This, in turn, facilitates a better understanding of AI systems and provides a basis for the development of interoperable and harmonized transparency-related standards.
This standard is intended for use by organizations of all types using any type of AI system.
The standard’s contents are as follows:
1 Scope
2 Normative References
3 Terms and Definitions
4 Abbreviations
5 Overview
6 Stakeholder Needs and Transparency Goals
7 Context Taxonomy
8 System Taxonomy
9 Model Taxonomy
10 Dataset Taxonomy
Appendix A (informative): Examples of Stakeholder Roles in Transparency Disclosure
Bibliography
Source: Rosstandart website