The Council agrees on a position to simplify the rules on artificial intelligence.

On 13 March, the Council agreed on its position regarding the proposal to simplify certain rules relating to artificial intelligence (AI).

The proposal forms part of the so-called‘Omnibus VII’legislative packageunder theEU’ssimplificationagenda. This package includes proposals for two regulations aimed at simplifying the EU’s digital legislative framework and the application of harmonised rules on artificial intelligence.

The Commission proposed extending thedeadline for the application of the rules on high-risk AI systemsby up to 16 months, so that these rules would enter into force once the Commission confirms the availability of the necessary standards and tools. It also proposedspecific amendmentsto the AI Act that would extend certainregulatoryexemptionsgranted to SMEs to small and medium-sized enterprises (SMEs), reduce requirements in a very limited number of cases, expand the scope for processing sensitive personal data for the detection and mitigation of bias,strengthen the powers of the AI Office, and reduce fragmentation in governance.

Main amendments introduced by the Council

The Presidency has treated the proposal asa top priority. Member States shared this sense of urgency and, with this in mind, generally maintained the main thrust of the Commission’s proposal.

However, the Council’s mandate adds a new provision to the AI Act prohibiting AI practices relating to thegeneration of non-consensual sexual and intimate content orchild sexual abuse material. The text also sets a fixed timetable for the deferred application of the high-risk rules: the new application dates wouldbe 2 December 2027for standalone high-risk AI systems and2 August 2028for high-risk AI systems embedded in products.

Furthermore, the Council’s mandate reinstates theobligation for providers to register AI systemsin the EU database for high-risk systems, where they consider their systems to be exempt from being classified as high-risk. It also reinstates thecriterion of strict necessityfor the processing of special categories of personal data in order to ensure the detection and correction of bias.

Furthermore, the textextendsthe deadline for the creation of regulatory AI testing environments by competent national authoritiesuntil2 December 2027. It also clarifies thepowers of the AI Officeregarding the supervision of AI systems based on general-purpose AI models, where both the model and the system are developed by the same provider, listingthe exceptionswhere national authorities remain competent, including law enforcement, border management, judicial authorities and financial institutions.

Finally, the Council’s mandate adds a newobligation for the Commission to provide guidanceto help economic operators of high-risk AI systems covered by sectoral harmonisation legislation comply with the high-risk requirements of the AI Act in a way that minimises the administrative burden.

Next steps

Following today’s adoption of the Council’s mandate, the Presidency will begin negotiations with the European Parliament.

Background

In October 2024, the European Council urged all EU institutions, Member States and stakeholders to prioritise work addressing the challenges identified in the reports by Enrico Letta (‘Much more than a market’) and Mario Draghi (‘The future of European competitiveness’). The Budapest Declaration of 8 November 2024 subsequently called for ‘launching a simplification revolution’, ensuring a clear, simple and smart regulatory framework for businesses and drastically reducing administrative, regulatory and reporting burdens, in particular for SMEs.

Since February 2025, following up on the call by EU leaders at that and subsequent meetings, the Commission has presented ten omnibus packages aimed at simplifying existing legislation on sustainability, investment, agriculture, small and medium-sized enterprises, digitalisation and common specifications, defence readiness, chemicals, digital issues including AI, the environment, the automotive sector, and food and feed safety. On 19 November 2025, the Commission presented its seventh omnibus package: the Digital Omnibus, comprising two proposals: one aimed at simplifying the EU’s digital legislative framework and the other at simplifying the application of harmonised rules on AI. This initiative stems from a broader policy objective: to improve the EU’s competitiveness by reducing the administrative burden on businesses and creating more favourable conditions for them to operate in the EU.

Further information: Council of Europe