The Dutch data protection authority has created a dedicated AI and algorithm supervision directorate, but it does not yet have formal AI Act powers

Key takeaways

• The notice is organisational. It creates no new duty and sets no compliance date.
• The AP says it has power in AI market surveillance only once the Dutch implementation act is in force.
• The AP says the deadline for designating AI supervisors was 2 August 2025, and the Dutch act is still being worked on.
• The AP points out that some AI Act rules already apply, including bans and transparency duties.

What did the AP announce on 9 October 2026?

The AP announced the start of a new directorate, Toezicht en Coördinatie AI/Algoritmes, or TCA. According to the notice, the directorate continues the coordinating supervision of AI and algorithms that the AP has carried out since 2023. It also prepares AI market surveillance, and it works with the Rijksinspectie Digitale Infrastructuur (RDI) on the "AI sandbox" and a joint AI Coördinatiecentrum. The AP says the start of TCA is a next step in preparing for the central role the cabinet plans to give the AP and the RDI in Dutch supervision of the AI Act.

Does the AP have AI Act enforcement powers yet?

No. The AP states: "De AP is beoogd markttoezichthouder op AI, maar nog niet aangewezen." It adds: "De AP beschikt pas over bevoegdheden in het AI-markttoezicht zodra de Nederlandse Uitvoeringswet AI-verordening van kracht is." The notice says talks on the AP's designation, powers and funding are under way at several levels, and that the AP reviewed the implementation act earlier this year and mapped the capacity and funding it needs. It also says the Netherlands is still working on the act, and that the deadline for designating supervisors was 2 August 2025. See also our earlier article on Dutch implementation.

How is TCA organised, and what do ARC and MTA do?

TCA has two departments, according to the notice. The department AI/Algoritmes Risicoanalyse & Coördinatie (ARC) looks at the impact of AI and algorithms on people, fundamental rights and society, and pays particular attention to high-impact algorithmic systems, algorithm registers and audit frameworks. The notice gives risk-selection instruments as an example, to be checked for discrimination and accuracy. The department Markttoezicht AI (MTA) prepares future supervision under the AI Act, builds technical expertise to investigate and test AI systems, and works on guidance and support for responsible innovation. For AI safety, the notice says national supervisors work with the AI Office, including on testing and evaluating frontier models and systems.

What does it mean for companies and public bodies?

The notice does not add a duty, but it says rules already apply. It says some AI uses are already banned and that transparency duties apply to AI-generated content such as video, images or text. It also says a ban on AI that virtually undresses people without consent applies from 2 December; the notice gives no year. For public bodies, the notice lists algorithm registers and audit frameworks among ARC's areas of attention; it sets no inspection plan. The notice says the cabinet plans to give the AP, together with the RDI, a central role, and that the AP is not yet designated.

What should compliance teams do now?

Treat the notice as a signal about who will supervise, not as a change in obligations.

• Check your AI Act obligations map for the bans and transparency duties the notice says already apply.
• Note the AP and the RDI as the expected supervisors, and mark the designation as pending until the Dutch implementation act is in force.
• If you hold or publish an algorithm register, confirm it is current, since ARC lists registers as an area of attention.
• If you are building a high-risk system, ask whether the AI sandbox the AP and RDI describe is open to you; the notice gives no application terms.

Which dates does the notice mention?

What we did not verify

What we opened: the AP news notice "AP zet volgende stap in toezicht op AI", dated 9 October 2026, on 25 October 2026 (UTC).

What we did not open: the PDF "Toezicht en coordinatie AI en algoritmes" linked from the notice, the Uitvoeringswet AI-verordening, the AP's earlier review of it, the AP notice of 20 April 2026, and any press coverage of the announcement.

What we refuse to claim: that the AP has any AI Act enforcement power today, a date for the implementation act, who the final designated authorities will be, or any year for the 2 December ban beyond what the notice says.

Related briefings

• AI Regulation News hub
• Europe AI regulation tracker
• How the Netherlands is implementing the AI Act with decentralised supervision
• Dutch AP and the AI transparency code deadline
• Dutch AP guidance on fundamental rights impact assessments and the national pilot
• AI Act Article 5 ban on AI used for intimate image abuse