European Union member states and the European Parliament have reached a provisional agreement to streamline the bloc's artificial intelligence regulations, part of a broader omnibus package aimed at reducing bureaucratic burdens on businesses. The deal, announced on Thursday, grants additional time for companies to comply with certain obligations under the EU Artificial Intelligence Act, while introducing a ban on AI-generated non-consensual explicit content.
The Digital Omnibus on AI, proposed five months ago, seeks to boost Europe's competitiveness in the AI sector by clarifying overlapping rules. Under the current framework, many firms face confusion over whether to follow the AI Act or sector-specific legislation, such as those governing machinery or medical devices. Swedish MEP Arba Kokalari, rapporteur for the Parliament's Internal Market committee, emphasized that the changes are not about weakening safety standards. 'We are not weakening any safety rules; we are clarifying the rules for companies in Europe,' she said. 'The current state is that companies are confused about whether they should follow the AI act or sectoral legislation … companies should not be regulated twice for one thing.'
Key Amendments and Timelines
The EU AI Act categorizes AI systems by risk level, from minimal to unacceptable, with stricter obligations for higher-risk applications. High-risk systems—those used in critical infrastructure, education, employment, migration, and border control—now have until December 2027 to comply, an extension of one year. For AI embedded in products like lifts or toys, the deadline is pushed to August 2, 2028. Small and medium-sized enterprises will benefit from simplified rules to avoid duplication between sectoral and AI requirements, according to the European Commission.
The agreement also establishes an 'EU-level sandbox' where developers can test AI products before market entry, a move designed to foster innovation without compromising oversight. This aligns with broader efforts to position Europe as a leader in responsible AI development, as noted in recent trends where Ireland, Germany, and the Netherlands lead Europe's AI talent surge.
Ban on Nudification Apps
A notable addition is the prohibition of AI systems that generate non-consensual sexually explicit content, including so-called 'nudification apps' that digitally remove clothing from images. The ban covers explicit images, videos, or audio created without consent, and applies to content depicting real humans—not synthetic AI characters. Companies must comply by December 2, implementing mandatory watermarking of AI-generated material. Renew Europe lawmaker Michael McNamara clarified that the rules target any content exposing a person's 'intimate parts' without permission, a response to incidents like Elon Musk's Grok chatbot generating explicit images of women and children.
Kokalari stressed the importance of clarity: 'We wanted to have clarity on what we think about [nudification apps] in Europe and that we are not accepting of it.' The measure reflects growing concerns over AI misuse, though it stops short of regulating fully synthetic content.
The provisional deal now awaits formal approval from the Council of the EU and the Parliament. Critics argue the delays and simplifications risk watering down protections, but proponents see them as necessary to avoid stifling innovation. As the EU navigates its digital agenda, similar tensions have emerged in other areas, such as trade talks with the US, where the bloc insists on maintaining its tech rules.
