Mandatory requirements for high-risk AI systems in EU postponed until end of 2027

The European Union (EU) is experiencing significant difficulties in establishing a regulatory framework for artificial intelligence, leading to the postponement of key obligations for market participants to later dates, Andres Ojaver, data privacy legal consultant at Hedman and lecturer on AI and personal data protection at Tallinn University of Technology, stated at the 8th Summit of IT Leaders of State Institutions in Baku, Report informs.

The specialist stressed that the implementation process of the European regulatory system is still far from complete. In particular, the effective dates for mandatory requirements for standalone high-risk AI systems, including solutions used in public services, have been pushed back from August 2026 to December 2027. Regulations for AI integrated into regulated products have been delayed until August 2028, while updated prohibited practices will take effect in December 2026.

Ojaver clarified that the reason for the postponement is the unreadiness of the supporting infrastructure, as well as the protracted process of developing unified harmonized standards. In addition, the European Commission continues to issue guidelines, while EU member states are still at the stage of establishing national market surveillance authorities.

Additional difficulties are created by the need to align the provisions of the AI Act with existing regulations such as GDPR, NIS2, the Cyber Resilience Act, as well as sector-specific regulatory acts.