The Council has agreed its position on a proposal to streamline rules on artificial intelligence, as part of the “Omnibus VII” legislative package within the EU’s broader simplification agenda. The proposal aims to bring greater legal certainty, more proportionate rules and more harmonised implementation of the AI Act across member states, while supporting innovation and EU competitiveness. The proposal for the regulation is available at this link.

Key elements include adjusting the timeline for applying high-risk AI system rules to ensure standards and tools are ready, extending regulatory exemptions to small mid-caps, and reinforcing the AI Office’s powers. The Council also added a new prohibition on AI practices generating non-consensual sexual content or child sexual abuse material and set fixed deadlines: December 2027 for stand-alone high-risk AI systems and August 2028 for those embedded in products. Additionally, the Council reinstated obligations for providers to register systems in the EU high-risk database and maintained the strict necessity standard for processing sensitive personal data for bias detection.

The proposal is part of a broader simplification revolution called for by EU leaders, responding to challenges identified in the Letta and Draghi reports on competitiveness. Since February 2025, the Commission has put forward ten Omnibus packages aimed at reducing administrative and regulatory burdens, particularly for SMEs.

Following the Council’s mandate, the presidency will now start negotiations with the European Parliament.