AI Act implementation delayed for certain high-risk systems

The European Union's Artificial Intelligence Act (AI Act) has seen a partial delay in its implementation, specifically concerning requirements for high-risk systems. While the act entered into force on August 1, 2024, its application is being rolled out in stages.

General provisions, prohibitions on unacceptable practices, and AI governance rules are scheduled to take effect from February 2, 2025. Subsequently, regulations for general-purpose AI models, management, and penalties will apply from August 2, 2025. However, the more detailed requirements outlined in Chapter III of the AI Act, related to classification, requirements, and obligations for high-risk systems, have been shifted.

These specific requirements will now apply from December 2, 2027, for systems used in employment, education, essential services, justice, or migration. A further deadline of August 2, 2028, is set for systems linked to products covered by EU harmonizing legislation. Notably, the 'Digital Omnibus on AI' regulation did not establish a general moratorium; transparency obligations remain on track.