Council of the EU approves Digital Omnibus on AI

Council of the EU approves Digital Omnibus on AI

On June 29 2026, the Council of the European Union (“EU”) announced its final approval of the Digital Omnibus on AI (“AI Omnibus”): a new regulation aimed at streamlining and simplifying certain rules on AI. With this approval, European institutions look to strengthen legal certainty and achieve a more harmonized application of AI rules across the single market, while preserving fundamental rights against certain abusive uses of this technology.

The reform is part of the Omnibus VII package, one of the components of the EU’s regulatory simplification agenda, and include proposals for two regulations aiming to simplify the digital regulatory framework and harmonize the application of Regulation (EU) 2024/1689 on AI (“AI Act”).

While the final approved text is pending publication, the Council has announced the main issues that the AI Omnibus will introduce:

Postponement of obligations applicable to high-risk AI systems

The amendment with the greatest practical impact is the postponement of the obligations regarding high-risk AI systems. These obligations were due to start applying on August 2, 2026, a date that economic operators viewed with concern due to the lack of sufficient standards and technical guidelines.

Under the newly approved timeline, the obligations will apply from December 2, 2027, for stand-alone high-risk AI systems (those listed in Annex III), and from August 2, 2028, for high-risk AI systems subject to the Union harmonization legislation on product safety listed in Annex I. This postponement aims to provide operators and authorities with greater predictability and time to adapt.

Prohibition of sexual deepfakes and AI-generated child sexual abuse material (“CSAM”)

The amendment adds a new provision to the AI Act prohibiting AI practices regarding the generation of non-consensual sexual and intimate content, as well as child sexual abuse material (“CSAM”). In particular, AI systems that generate nude images of real people or manipulate existing images to remove clothing and reveal intimate body parts will be prohibited.

This ban is due to apply from December 2026, making it one of the most immediate measures in the package. It also reflects the commitment to ensuring that technological progress goes hand in hand with protecting fundamental values.

Deadlines regarding AI regulatory sandboxes and transparency solutions

The AI Omnibus also adjusts other important deadlines: (i) it postpones the deadline for competent national authorities to establish AI regulatory sandboxes to August 2, 2027; and (ii) it reduces the grace period for providers to implement transparency solutions for artificially generated content from six months to three months, with the new deadline set for December 2, 2026.

Competences of the AI Office

At an institutional level, the text clarifies any doubts regarding the division of competences. It confirms that the AI Office supervises AI systems based on general-purpose AI models, where the same provider develops the model and the system. However, it reserves a role  for national authorities in areas including law enforcement, border management, judicial authorities and financial institutions.

Interaction with sectoral legislation

Another matter addressed is the interaction of AI rules with sectoral legislation (in areas such as medical devices, toys, elevators, and watercraft), and any possible overlaps between the high-risk requirements from the AI Act and those from sectoral legislation. To avoid overlap, the AI Omnibus provides for a mechanism that, through delegated acts, allows the application of specific requirements or obligations under the AI Act to be limited where the applicable sectoral legislation provides an equivalent level of protection.

Furthermore, the products covered by Regulation (EU) 2023/1230 on Machinery are exempt from direct applicability of the AI Act. The European Commission (the “Commission”) is empowered to adopt delegated acts under the Machinery Regulation , which would add health and safety requirements with respect to high-risk AI systems.

Finally, a new obligation has been introduced requiring the Commission to provide guidance to economic operators of high-risk AI systems covered by sectoral harmonization legislation, on how to comply with the requirements of the AI Act while minimizing the compliance burden.

The legislative act will be published in the coming days in the Official Journal of the EU and will enter into force three days after official publication.