Cyprus Presidency Secures EU Deal to Simplify AI Rules
EP and Council Under Cyprus EU Presidency Reach Provisional Agreement on Simplifying AI Rules
The Cyprus Presidency of the Council of the EU and European Parliament negotiators reached a provisional agreement on simplifying and streamlining certain rules governing artificial intelligence, under the European Union’s “Omnibus VII” legislative package for the simplification agenda, following negotiations in Brussels early Thursday morning. Cyprus’ Deputy Minister for European Affairs, Marilena Raouna, described the agreement as “clear evidence of the ability of the European institutions to act swiftly and deliver on their commitments.”
“The agreement significantly supports our companies by reducing recurring administrative costs,” Raouna said, adding that the provisional deal ensures “legal certainty and a smoother and more harmonised implementation of the rules across the Union.” According to Raouna, the agreement strengthens “the EU’s digital sovereignty and competitiveness, while at the same time reinforcing the protection of children against risks linked to AI systems.” She added that the agreement represents the first deliverable under the “One Europe, One Market” roadmap.
For her part, European Commission President Ursula von der Leyen welcomed the agreement on the “Digital Omnibus on AI,” saying in a social media post that it creates “a simple and innovation-friendly environment for our European AI ecosystem to grow.” As the President noted, “at the same time, we are strengthening protections for our citizens. For safe and simple AI governance in Europe.”
The co-legislators added a new provision to the AI Act prohibiting AI practices involving the generation of non-consensual sexual or intimate content, as well as child sexual abuse material (CSAM). The provisional agreement also sets a clear timeline for the delayed application of rules concerning high-risk AI systems. The new rules will apply from 2 December 2027 for standalone high-risk AI systems and from 2 August 2028 for high-risk AI systems embedded in products.
In addition, providers will once again be required to register high-risk AI systems in the relevant EU database, even in cases where they believe their systems are exempt from such classification. The agreement also reinstates the requirement of “strict necessity” for the processing of special categories of personal data for the purpose of detecting and correcting potential bias.
The agreement further postpones until 2 August 2027 the deadline for establishing national AI regulatory sandboxes, while reducing from six months to three months the adjustment period granted to providers for implementing transparency measures for AI-generated content. The new deadline is set for 2 December 2026.
Furthermore, the competences of the AI Office are clarified regarding the supervision of AI systems based on general-purpose AI models where both the model and the system are developed by the same provider. At the same time, the agreement specifies areas where national authorities will remain competent, including law enforcement, border management, judicial authorities and financial institutions.
As regards the industrial use of artificial intelligence and the interaction between the AI Act and sector-specific legislation in areas such as medical devices, toys, lifts, machinery and watercraft, the co-legislators agreed on a mechanism aimed at avoiding overlaps between different regulatory requirements. Through implementing acts, the application of the AI Act may be limited in cases where sectoral legislation already includes equivalent AI-related requirements.
The agreement also exempts the Machinery Regulation from the direct application of the AI Act. In parallel, the European Commission is empowered to adopt delegated acts under the Machinery Regulation in order to introduce additional health and safety requirements for AI systems classified as high-risk.
Finally, the agreement foresees that the European Commission will issue guidelines to assist businesses and economic operators active in sectors covered by specific legislation in complying with the AI Act requirements while minimising administrative burdens.
The provisional agreement must now be formally endorsed by both the Council of the EU and the European Parliament before undergoing legal and linguistic revision ahead of its final adoption in the coming weeks.