European Union member states are advancing a compromise that would expand the authority of the bloc's central artificial intelligence regulator while simultaneously imposing new limits on its enforcement powers, according to different sources online.
The "second compromise" text, dated early February according to Euractiv, represents the latest effort to fine-tune the EU's AI Act as the law approaches a critical implementation phase in August. The revisions come amid broader negotiations over the so-called Digital Omnibus package, which seeks to streamline the bloc's growing web of technology regulations.
At the heart of the proposal is an expanded role for the European AI Office, a relatively new body within the European Commission tasked with overseeing advanced AI systems. By channelling more enforcement authority through Brussels, officials hope to create a more predictable regulatory environment for companies operating across the single market.
Tech lobbying group CCIA Europe and Digital Europe have supported centralising certain enforcement in the AI Office for consistency but criticised the broader Omnibus package for lacking ambition, arguing it fails to fully address overlaps, asymmetric rules favouring smaller companies, and technical feasibility issues that continue to burden innovators.
The compromise builds on a broader regulatory package the European Commission proposed last November. If adopted, it could delay certain fines and ease some compliance requirements for high-risk AI applications (with key obligations now phased toward full application by August 2026), giving companies more time to adapt to the new rules.
Negotiators from the Commission, member states and the European Parliament are expected to enter intensive talks by spring to finalise the changes. The outcome will shape how Europe regulates AI systems as other major economies, including the U.S. and China, develop their own approaches to the technology.
