EU AI Act Enforcement

The European Union’s AI Act is moving from a rulebook to active policing, and the Netherlands plans to split that job across 10 regulators. (pinsentmasons.com) The Dutch government opened a consultation this week on draft legislation that assigns AI Act oversight by sector, with comments due by June 1, 2026. Pinsent Masons said the plan comes more than eight months after the August 2, 2025 deadline for member states to put national enforcement laws in place. (pinsentmasons.com) Under the draft, the Dutch Data Protection Authority and the State Inspectorate of Digital Infrastructure would act as coordinating authorities, while sector bodies such as the Health and Youth Care Inspectorate, the food and product safety authority, and the labour inspectorate would supervise AI inside their own domains. (pinsentmasons.com) The AI Act itself entered into force on August 1, 2024, but its obligations arrive in stages rather than all at once. The European Commission said the law bans some uses outright, imposes transparency rules on systems such as chatbots, and sets stricter controls for high-risk tools such as medical software and recruitment systems. (commission.europa.eu) In the Netherlands, the Dutch Data Protection Authority said the first requirements started applying from February 2025, including bans on certain AI systems and a duty for organisations using AI to ensure staff have enough AI literacy. The Commission’s guidance says obligations for providers of general-purpose AI models started applying on August 2, 2025. (autoriteitpersoonsgegevens.nl) (digital-strategy.ec.europa.eu) The biggest compliance wave is still ahead. The European Parliament’s research service said the AI Act’s general date of application, including enforcement rules in Chapter 9, is August 2, 2026, and the full roll-out is foreseen by August 2, 2027. (europarl.europa.eu) (digital-strategy.ec.europa.eu) That timetable helps explain why national regulator maps matter now. A company selling one high-risk AI system into hospitals, workplaces and consumer products could face different Dutch supervisors depending on where and how the system is used under the sector-based model described in the draft. (pinsentmasons.com) (commission.europa.eu) The Netherlands has been building that structure for several years. The Dutch Data Protection Authority says it has served since 2023 as the coordinating market surveillance authority for algorithms and AI involving risks to fundamental rights, through a dedicated Department for the Coordination of Algorithmic Oversight. (autoriteitpersoonsgegevens.nl) Brussels is also trying to make the law easier to run before the hardest parts bite. The European Commission published a “Digital Omnibus on AI” proposal on November 19, 2025, describing it as a package of targeted simplification measures for the AI Act. (digital-strategy.ec.europa.eu) For companies building or embedding high-risk AI, that leaves two clocks running at once: the European Union is still refining the system, and national governments are deciding who will enforce it. In the Netherlands, that answer now looks less like one watchdog and more like a network. (digital-strategy.ec.europa.eu) (pinsentmasons.com)