DWP artificial intelligence Universal Credit benefit reviews, Digimart Global (United Kingdom): The Department for Work and Pensions (DWP) is trialling artificial intelligence to help decide which Universal Credit claimants should face intrusive benefit reviews. According to the department’s permanent secretary, a live test of a machine learning model is currently in use for Targeted Case Reviews (TCRs), a programme that reassesses more than one million claims every year.
In a letter to Parliament’s Public Accounts Committee, Dame Sarah Healey said the DWP is running a “limited live test” of the model to support agents in choosing which cases to target for review. She stressed that “human consideration remains central” and said the model does not make decisions on whether a review proceeds, benefit entitlement, payment changes, suspensions or claim closures.
The department says the anti-fraud initiative is set to save £13bn by the end of the decade. However, researchers from the Administrative Fairness Lab, reporting earlier this year, found that only 21 per cent of reviews identify any incorrect payment, and warned that hundreds of thousands of people are subjected to the process unnecessarily.
Mark Bennett of the University of Liverpool and the Administrative Fairness Lab described the use of machine learning as a “potential step change” that “comes with clear risks”. He pointed to a separate AI tool the DWP previously used to screen Universal Credit advance payments, which, he said, disproportionately flagged older, disabled and non-UK claimants for fraud checks.
Dame Sarah said the trial operates under the usual safeguards and equality standards, and that the DWP would “consider alternative options” if it fails to deliver as expected. She confirmed that the specific criteria used to flag cases and other features of the model must remain confidential to protect the “integrity of the benefit system”.
The Administrative Fairness Lab research also found that many claimants selected for TCRs are repeatedly reviewed, often only months apart, causing what researchers described as “feelings of criminalisation” and exacerbating anxiety. Mr Bennett said the scale of the scheme means the impact of any mistakes is likely to be widely felt.









