10 to the 23 AI logo

Domain Atlas / Public benefits & eligibility

Case fileRotterdam, Netherlandslarge deployment

Rotterdam welfare-fraud risk model

An independent audit of Rotterdam's welfare-fraud risk model documented scores skewed against already-vulnerable groups, and the city suspended the system's use following the scrutiny.[4]

What happened

Rotterdam used a machine-learning model (originally built with Accenture) to score welfare recipients for fraud-investigation priority. Journalists at Lighthouse Reports and partners obtained the model and training data, and their audit documented that risk scores skewed against already-vulnerable groups — women, parents, people with limited Dutch — with individual features like language skill acting as proxies. The city suspended the system's use following the scrutiny.

The sociotechnical reading

Rotterdam is the Atlas's clearest case of the audit-as-actor: the control that finally bound the system came from investigative access to the model itself, which is exactly what vendor opacity usually prevents. It also illustrates selection harms that live outside output accuracy — the model decided who got investigated, and investigation is itself a burden. Governance that only measures "was the flag correct" misses the question "who bears the process". The equity-measurement patterns in the Practice Library exist for this distinction.

The concepts used in this reading are defined in the Field Guide; the governance responses live in the Practice Library.

Grounding sources for this case

The same sources that ground this model organization in the PAN library — evaluations, government documents, investigative reporting, and advocacy documentation, each labeled by tier.

wiredlighthousereports2023GroundingInvestigative

WIRED / Lighthouse Reports, Inside the suspicion machine (2023) https://www.wired.com/story/welfare-state-algorithms/

https://www.wired.com/story/welfare-state-algorithms/

Appears in: PAN framework development

Grounds: deployment audit: Rotterdam welfare-fraud algorithm

Seeing your organization in this case file?

The histories here are documented after the harm. Mapping a live deployment's pathways and pressures — before the incident report — is engagement work: intake, diagnosis, prescription, and monitoring, with every limitation stated.

Sources & Evidence

Claims made on this page and what supports them. The full registry lives in Evidence.

EmpiricalAn independent audit of Rotterdam's welfare-fraud risk model documented scores skewed against already-vulnerab…

An independent audit of Rotterdam's welfare-fraud risk model documented scores skewed against already-vulnerable groups, and the city suspended the system's use following the scrutiny.

wiredlighthousereports2023GroundingInvestigative

WIRED / Lighthouse Reports, Inside the suspicion machine (2023) https://www.wired.com/story/welfare-state-algorithms/

https://www.wired.com/story/welfare-state-algorithms/

Appears in: PAN framework development

Grounds: deployment audit: Rotterdam welfare-fraud algorithm