The Netherlands, explained in English · Thursday 29 October 2026Thu 29 Oct

ENTR

Society

Collection agencies use more AI, risking wrong decisions

The Dutch data protection authority (Autoriteit Persoonsgegevens, AP) says collection agencies increasingly use algorithms and AI, and that decisions about people in debt must never be left entirely to a computer.

Collection agencies are using algorithms and AI more and more, according to a sample survey of 29 agencies by the Autoriteit Persoonsgegevens (AP). The AP says the work goes faster, but the technology brings risks of incorrect decisions, discrimination and unequal treatment. It has issued recommendations to the sector.

For anyone facing a collection case, the AP says, the consequences can be large. Agencies must never leave decisions about debtors entirely to the computer, it says, and a human employee must be involved in important decisions.

Many agencies use AI to estimate the chance that a debt can be collected, based on variables including a person's credit score, location and payment history. Staff then use the result to decide which debtors to approach first.

Agencies also use the systems to "enrich" debtor files. The software searches registers or the internet for contact details, bank details, company details or social media activity, and adds them to the file.

The AP found that agencies do not always know the rules on using algorithms and AI. It considers it a realistic possibility that fully automated, and therefore prohibited, decision-making happens in practice without agencies realising it.

AI can also make mistakes. That is why it is important that it is clear and transparent how a decision is reached, and that a flesh-and-blood collection employee can account for it. Otherwise you are completely in the dark while AI decides about you.Katja Mur, AP board member

What happens next

  • The AP is holding a round-table meeting with collection sector organisations on 2 December 2026 to discuss its findings and recommendations.
  • Organisations that want to take part must register before 11 November 2026.