Russia Starts Work on National AI Security Standard
The Federal Service for Technical and Export Control (FSTEC) is drafting Russia’s first dedicated standard for the safe development of artificial intelligence systems, aiming to address AI-specific cyberthreats.
The document, expected to be presented by the end of 2025, will focus on eliminating unique cybersecurity risks inherent to AI technologies.
Vitaly Lyutikov, FSTEC’s First Deputy Director, confirmed the initiative during the BIS-2025 forum, adding that the draft will be released for public discussion in the coming months.
The new standard will not replace but rather complement the existing “Secure Software Development Requirements.” As Igor Bederov, head of investigations at T.Hunter, explained, the SDLC standard covers general practices, but AI systems have fundamental differences that require a specialized approach to threat modeling throughout their lifecycle.
The framework will establish methodologies for building secure AI solutions, primarily for government information systems, though experts believe it will also guide the commercial sector.
The project’s presentation is scheduled for late 2025. According to Alexey Lukatsky, a security consultant at Positive Technologies, AI safety is already partially reflected in FSTEC’s new Order No. 117, which comes into force on March 1, 2026. However, while the order defines requirements, the new standard will provide detailed instructions on how to meet them.
Experts say the emergence of a national AI security standard will help systematize approaches to safeguarding advanced technologies right from the design stage.
