PublicAlgoBeta
All case studies

UK · December 2024

The DWP model that flags Universal Credit advance requests for fraud

Image: Big Brother Watch. All rights reserved by the owner.

Reported by
The Guardian, Public Law Project, Big Brother Watch, Computer Weekly
Story language
EN
Methods
  • FOI / records request
  • Data analysis

A records release forced the UK's Department for Work and Pensions to admit statistically significant disparities in a fraud tool that scores Universal Credit advance requests. Its own later figures show non-UK nationals were referred about 2.27 times as often as UK nationals in 2024/25, and about 3.35 times in 2025/26. DWP says a caseworker decides every flagged case. It published its fairness assessment and committed to retrain the model, which stays in use until a retrained version is tested.

computerweekly.comThe DWP model that flags Universal Credit advance requests for fraudA supervised machine-learning model used by the UK Department for Work and Pensions (DWP). It scores each Universal Credit advance request for fraud risk before payment, and high scorers are referred to a caseworker, who approves or declines.United Kingdom · Reported by The Guardian, Public Law Project, Big Brother Watch and others (opens in a new tab)

How they investigated

  1. FOI / records request

    The Public Law Project obtained DWP's internal 11-page fairness analysis, dated February 2024, through a freedom of information request. Separately, a Big Brother Watch researcher asked for DWP's data protection impact assessments. DWP released redacted copies and withheld the rest. The Information Commissioner backed the refusal, and the researcher appealed to a tribunal, which heard the case on 30 May 2024.

  2. Data analysis

    The statistics are DWP's own, not the journalists'. The internal analysis tested referral and outcome disparities by age, disability, marital status and nationality. DWP later published the same kind of analysis for 2024/25 and 2025/26, comparing how often a group is referred with how often those referrals prove correct. The contribution of journalists and campaigners was to obtain, read and challenge it.

Read alongside