Artificial Intelligence · Denmark
Artificial Intelligence - Denmark
As an EU member state, Denmark applies the EU AI Act directly, with its risk-based obligations phasing in (prohibited-use bans from February 2025, GPAI rules from August 2025, and the bulk of high-risk obligations from August 2026). Denmark was among the first member states to pass national implementing legislation — Act No. 467 of 14 May 2025 — designating competent authorities, market-surveillance powers and penalties. The Agency for Digital Government coordinates oversight, supported by the Data Protection Authority and the Court Administration, plus a joint regulatory sandbox.
The EU AI Act (Regulation (EU) 2024/1689) is directly applicable in Denmark, setting a risk-based framework with prohibited practices, high-risk system obligations, GPAI/transparency rules and staggered application dates running through 2026-2027.
Act No. 467 of 14 May 2025 supplements the AI Act in Danish law, designating national competent authorities and laying down market-surveillance powers and penalties; it entered into force on 2 August 2025, making Denmark one of the first member states to legislate.
The Agency for Digital Government (Digitaliseringsstyrelsen) is the notifying authority, a market-surveillance authority, and Denmark's single point of contact toward the European Commission and other member states under Article 70 of the AI Act.
Market-surveillance responsibilities are shared: the Agency for Digital Government (general), the Danish Data Protection Authority (Datatilsynet) and the Danish Court Administration (Domstolsstyrelsen) hold distinct remits, with powers to demand information, order compliance, impose temporary bans and conduct inspections.
Datatilsynet and Digitaliseringsstyrelsen jointly run a free regulatory sandbox offering expert guidance on GDPR and AI Act risk classification; it gives no exemptions or pre-approval. A first round concluded in October 2025 and a second round opened for applications.
The Agency for Digital Government has published practical guides for public authorities and private companies on responsible use of generative AI, complementing the binding AI Act framework.
Machine-assisted translation · verified 5/23/2026 · orientation, not legal advice. English version →