Dutch tax agency fraud algorithm discriminated against dual nationals

A systemic failure in the Dutch tax authority's fraud-detection algorithms led to discriminatory targeting of dual nationals, causing thousands of families to be wrongly accused and face financial hardship; the event attracted regulatory scrutiny and political repercussions in 2024. The AP AI & Algorithmic Risks Report formally acknowledges systemic AI risks linked to this case.

Dutch Tax and Customs Administration (Belastingdienst) · Incident Jan 1, 2024 · Indexed Jun 8, 2026 · 3 sources

The use of dual nationality as a risk indicator transformed a fraud-detection tool into a mechanism for systemic discrimination.
What
A systemic failure in the Dutch tax authority's fraud-detection algorithms led to discriminatory targeting of dual nationals, causing thousands of families to be wrongly accused and face financial hardship; the event attracted regulatory scrutiny and political repercussions in 2024.
Incident date
Jan 1, 2024
Who
Dutch Tax and Customs Administration (Belastingdienst)
Failure mode
Policy Violation
AI surface
Algorithmic Decision
Severity
Catastrophic

What happened

The Dutch Tax and Customs Administration deployed a risk-scoring system that disproportionately flagged citizens with dual nationalities as high-risk for fraud. This systemic failure led to thousands of families being wrongfully accused of fraud and forced to repay childcare benefit allowances they were entitled to. The incident caused extreme financial distress for affected citizens and contributed to political upheaval in 2021.

What broke inside the model

Failure path · this incident · Policy Violation
  1. 01 · TriggerParents claim childcare benefits; the tax agency's risk model scores them for fraud.
  2. 02 · Model stepThe model uses nationality as a proxy variable in suspicion scoring.
  3. 03 · Control gapIndividualized evidence is replaced by an algorithmic presumption of guilt; no review questions the proxy.
  4. 04 · FailureThousands of families, disproportionately dual nationals, are branded fraudsters and ordered to repay.
  5. 05 · ConsequenceFamilies are ruined, children enter care, and the Dutch government resigns over the scandal.

The system used discriminatory proxy variables, such as nationality, to automate fraud suspicion. This logic replaced individualized evidence with biased algorithmic presumptions of guilt, creating systemic discrimination.

Public visibilityHigh
Regulatory exposureActive
Customer impactClass-wide
Financial impactEstimated
Time to disclosureMonths
  1. PrimaryAI & Algorithmic Risks Report Netherlands - Winter 2023 2024autoriteitpersoonsgegevens.nl
  2. PressXenophobic machines: Discrimination through unregulated AIamnesty.org
  3. Court FilingDutch child benefit scandal: origin and latest developmentsec.europa.eu
Permalinkhttps://failureindex.ai/failures/dutch-tax-agency-fraud-algorithm-discriminated
CitationAI Failure Index. "Dutch tax agency fraud algorithm discriminated against dual nationals" (FI-0315). Realm Labs. https://failureindex.ai/failures/dutch-tax-agency-fraud-algorithm-discriminated (indexed Jun 8, 2026).
Share cardA branded image of this record for posts and slides.

Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0315. Full dataset at /data.

Note from Realm Labs, the Index steward

How Realm fits

Controls for this failure mode
  • Prism
  • OmniGuard

This entry sits in the index's predictive wing: a system that scores, ranks, perceives, or steers rather than generates. Realm's runtime layer is built for the generative and agentic systems now moving into these same decision seats, where it watches a model's internal state and holds an unsupported claim or an unchecked action before it commits. The control gap on this record, an automated decision that reached people with no runtime check in front of it, is the same gap. The index keeps predictive failures on the record because the pattern carries straight into the systems shipping today.