Council compromise text advances EU AI Act changes
The compromise text would amend AI Act rules on AI literacy, conformity assessment, sandboxes, and high-risk system obligations.
The Council of the European Union has confirmed agreement on a compromise text for the Digital Omnibus on AI, a proposal intended to simplify parts of the EU AI Act’s implementation while preserving protections for health, safety, and fundamental rights.
The Permanent Representatives Committee confirmed the agreement on 13 May 2026, following informal negotiations between the EU institutions on 6 May. The Council Presidency was authorised to send a letter to the European Parliament stating that, if Parliament adopts the text at first reading, the Council will approve Parliament’s position.
The compromise text amends Regulation (EU) 2024/1689 on AI and Regulation (EU) 2018/1139 on civil aviation. It says targeted changes are needed because delayed standards, national governance structures, and conformity assessment frameworks have created compliance burdens heavier than expected.
The proposal would adjust several AI Act implementation rules, including provisions on AI literacy, treatment of small mid-cap enterprises, conformity assessment, AI regulatory sandboxes, real-world testing, and the role of the AI Office. It would also simplify some registration and monitoring requirements while providing more time for high-risk AI obligations to apply.
One major addition concerns prohibited AI practices. The text would prohibit placing on the market, putting into service, or using AI systems that generate or manipulate realistic non-consensual intimate images, videos, audio, or similar material of identifiable people. It would also prohibit AI systems that generate or manipulate child sexual abuse material, subject to limited lawful exceptions.
The compromise text also modifies the AI literacy obligation. Instead of requiring providers and deployers to ensure a sufficient level of AI literacy among staff, the revised wording would require them to take measures to support AI literacy, while clarifying that they are not required to guarantee a specific level for each individual.
For high-risk AI systems, the compromise text proposes delayed application dates for certain obligations: 2 December 2027 for systems classified as high-risk under Article 6(2) and Annex III, and 2 August 2028 for systems classified as high-risk under Article 6(1) and Annex I. The text says this is intended to address implementation challenges linked to delayed standards, guidance, and national competent authorities.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!