On 29 June 2026, the Council of the European Union formally adopted the Digital Omnibus on AI, completing the legislative cycle after the European Parliament approved the amendments on 16 June 2026. The amendments modify Regulation (EU) 2024/1689 (the EU AI Act) and enter into force as directly applicable EU law across all member states.
Under Annex III of the AI Act, compliance deadlines for stand-alone high-risk AI systems used in employment, education, biometric identification, and critical infrastructure management shift from 2 August 2026 to 2 December 2027. For Annex I high-risk AI systems embedded in products subject to sector-specific legislation (medical devices, machinery, toys, and civil aviation components), the deadline extends to 2 August 2028. The synthetic content watermarking obligation under Article 50(2) for outputs generated before 2 August 2026 is deferred to 2 December 2026. AI regulatory sandbox establishment under Article 57 is postponed to 2 August 2027.
HR technology vendors, recruitment platforms, educational assessment tools, biometric identification system operators, and AI-enabled critical infrastructure managers gain an additional 16 months before they must satisfy the AI Act's conformity assessment, technical documentation, logging, transparency, and human oversight obligations. Integrators embedding AI capability into regulated Annex I products gain 24 months. Both categories must continue to comply with pre-existing product safety and sector-specific regulatory requirements during the extended period.
Two new prohibited practices are added under Article 5 of the AI Act, effective 2 December 2026: generating or manipulating non-consensual intimate imagery and child sexual abuse material. These prohibitions sit alongside the existing bans on social scoring and subliminal behavioural manipulation. No transitional provisions apply to the new prohibitions.
Licentium advises technology companies and regulated entities on EU AI Act compliance and implementation planning. Organisations assessing their position under the revised deadlines are invited to contact us. Work we undertake includes Annex I and Annex III high-risk system classification, conformity assessment preparation, technical documentation design, human oversight programme development, and multi-jurisdictional AI regulatory mapping.