16 EU states lack AI Act enforcers days before Dublin AI summit
What's the deal? Just days before the EU's first leaders' summit on artificial intelligence in Dublin, 16 member states have yet to fully designate the national authorities meant to enforce the AI Act. Six countries have named no authority at all, while 10 have only provisional ones.
Why now? The deadline to appoint enforcers passed in August 2025. The gap is now more pressing because the AI Act's first major enforcement provisions took effect in August 2026, putting national authorities in charge of overseeing many AI systems.
How enforcement splits: National authorities handle most AI systems on the ground, while the European Commission's AI Office oversees general-purpose models. The Commission has already begun enforcement against frontier-model providers, but that EU-level work does not replace national oversight — leaving implementation uneven across the bloc.
What's the endgame? The EU wants to regulate AI uniformly across 27 countries. The Dublin summit is meant to showcase that ambition, alongside a €30B plan for AI compute hubs — which has now slipped a second time.
What could go wrong? Without named enforcers, companies face patchy oversight and legal uncertainty about who regulates them where. Countries that miss designation deadlines can face infringement action from the Commission.
The strain runs deeper than one missed milestone. The EU has already delayed high-risk-system obligations to December 2027, citing a need for standards and support tools still under development.
The signal: The story exposes the distance between the EU's goal to set the global standard for AI rules and its capacity to enforce them consistently. As it races to catch the US and China on AI, Europe is struggling to finish its own homework.
Read more: Politico Europe
Image credit: DesignRecipe
