Regulator of algorithms established in the Netherlands

A new unit has been established within the Autoriteit Persoonsgegevens (the AP) to lead the work, which will involve identifying and analysing cross-sector and overarching risks and effects of algorithms, optimising existing work that is ongoing in relation to the oversight of algorithms across sectors and within government, and developing guidance and other common standards on the use of algorithms.

The Dutch government has confirmed that it intends to clarify the longer-term goals and activities of the algorithm regulator later this year. Legislative reform may be on the agenda if the government decides the AP’s existing powers should be expanded.

The AP is already responsible for overseeing compliance with data protection law in the Netherlands, which includes addressing unlawful processing of personal data via artificial intelligence (AI) systems. The AP is to receive additional government funding to fulfil its expanded brief.

Amsterdam-based technology law expert Nienke Kingma of Pinsent Masons said: “A focus area for the AP will be strengthening the supervision of algorithms that process personal data unlawfully, with particular attention to promotion of transparency and risks of discrimination. The upcoming AI Regulation, currently being negotiated by European legislators, will only strengthen these elements and provide further concrete instruments and concepts that can be built on.”

Alexandra van Huffelen, secretary of state in the Ministry of the Interior and Kingdom Relations in the Dutch government, said: “The establishment of an algorithm supervisor stems from the ambition to better protect public values and fundamental rights in the use of algorithms.”

The expansion of the AP’s brief to include oversight of algorithms comes after a review of government use of algorithms in the Netherlands found that there is often a lack of criteria to guide their development and purpose and a lack of controls in place to govern their use. Government bodies in the Netherlands are now obliged to disclose the algorithms they use via a new AI register.

Kingma said: “The register was initiated by the Dutch government as part of its broader aim of investigating and getting more control over algorithms in terms of transparency and discrimination. The establishment of the register was specifically triggered by the recent serious problems uncovered with the use of algorithms by governmental organisations, such as in the recent childcare benefits scandal, in which the government’s use of algorithms led to discrimination and privacy violations.”

Prominent legal experts have advised against tasking the AP with supervision on algorithms, signalling the risk that the AP might be too focused on personal data processing and insufficiently address other interests at stake, such as competition and consumer protection law. The Netherlands Scientific Council for Government Policy recommended tasking sectoral supervisory authorities with the enforcement of AI regulation instead.

Amsterdam-based technology law expert Wouter Seinen described the choice of the AP as the new regulator as “remarkable”.

Seinen said: “The risk of an over-focus on personal data processing and a lack of expertise in other areas of law is real, especially since the AP is already struggling with understaffing issues and openly complaining about insufficient financial resources and has cited this as the reason for over 10,000 complaints remaining unanswered. It’s a bit like adding another wing whilst the main house is on fire.”

“Getting to grips with this much broader remit and recruiting the requisite legal and domain experts will be a big challenge. Luckily, the AP has developed good practices in collaborating with other authorities and has indicated it intends to follow this approach for the supervision on algorithms too,” he said.