Council Agrees Position to Streamline Rules on Artificial Intelligence
Today, the Council reached a significant milestone by agreeing on its position regarding the proposal to streamline certain rules concerning artificial intelligence (AI). This proposal is part of the broader “Omnibus VII” legislative package within the EU’s simplification agenda, aiming to simplify the EU’s digital legislative framework and establish harmonised rules on AI.
Importance of Streamlining AI Rules
The streamlining of AI rules is deemed essential for ensuring the EU’s digital sovereignty. The presidency emphasized the urgency with which they approached this proposal, aiming for timely application of the AI Act. The proposal is designed to provide greater legal certainty, enhance proportionality in regulations, and ensure consistent implementation across member states.
Key Provisions of the Proposal
The Commission has proposed adjustments to the timeline for applying rules on high-risk AI systems, potentially delaying these rules by up to 16 months until the necessary standards and tools are confirmed. Additionally, amendments to the AI Act have been suggested to:
- Extend regulatory exemptions granted to small and medium-sized enterprises (SMEs) to include small mid-caps (SMCs).
- Reduce requirements in limited cases.
- Allow processing of sensitive personal data for bias detection and mitigation.
- Reinforce the powers of the AI Office.
- Reduce governance fragmentation.
New Provisions and Obligations
In its mandate, the Council has included a new provision that prohibits AI practices related to the generation of non-consensual sexual and intimate content and child sexual abuse material. It also establishes fixed timelines for the delayed application of high-risk rules, with new application dates set for:
- December 2, 2027, for stand-alone high-risk AI systems.
- August 2, 2028, for high-risk AI systems integrated into products.
Moreover, the Council reinstates the requirement for providers to register AI systems in the EU database for high-risk systems if they consider their systems exempt from classification. The standard of strict necessity for processing special categories of personal data for bias detection and correction has also been reinforced.
Regulatory Sandboxes and Guidance
The deadline for establishing AI regulatory sandboxes at the national level has been postponed until December 2, 2027. Additionally, the Council clarifies the competences of the AI Office for supervising AI systems based on general-purpose AI models, noting exceptions where national authorities retain competence, such as in law enforcement, border management, judicial authorities, and financial institutions.
Lastly, a new obligation has been introduced for the Commission to provide guidance aimed at assisting economic operators of high-risk AI systems in complying with the AI Act’s requirements while minimizing compliance burdens.
Next Steps
With the Council’s mandate approved, the presidency is set to commence negotiations with the European Parliament. This initiative is part of a larger movement within the EU, as highlighted in the October 2024 European Council meeting, which emphasized the importance of addressing challenges identified in key reports and launching a simplification revolution to create a clearer and more streamlined regulatory framework for businesses.
Since February 2025, following calls from EU leaders, the Commission has introduced ten ‘Omnibus’ packages designed to simplify existing legislation across various sectors, including sustainability, investment, agriculture, and notably, AI. The seventh omnibus package, submitted on November 19, 2025, focuses on simplifying the EU’s digital framework and the implementation of harmonised rules on AI, aligning with the overarching goal of enhancing EU competitiveness.
