DWP Tests AI To Select Universal Credit Reviews, Raising Concerns Over Transparency and Fairness
The department says AI will only help select cases for review, while researchers warn of potential disadvantages for some claimants
The Department for Work and Pensions (DWP) is testing an artificial intelligence system to help identify Universal Credit claims for Targeted Case Reviews, prompting concerns from researchers and campaigners over transparency and the use of automated tools in welfare decisions. The department said human staff remain responsible for decisions, with the technology used only to support case selection.
The confirmation follows the DWP's disclosure that it is trialling a machine learning model as part of its efforts to tackle fraud and incorrect payments. According to the department, the technology helps identify claims that may require further review rather than deciding whether a claimant should receive benefits, face a payment change or have their claim closed.
AI Trial Focuses on Universal Credit Reviews
The system is being tested within the Targeted Case Review programme, which examines Universal Credit claims. The department's permanent secretary, Dame Sarah Healey, said the AI model is designed to help officials prioritise cases for review while human judgement remains at the centre of the process.
The DWP said advanced analytics can improve fraud detection and reduce incorrect payments. A spokesperson said Targeted Case Reviews had already saved taxpayers £1.7 billion by preventing incorrect Universal Credit payments and added that decisions would continue to be made by DWP staff.
'Decisions on cases selected for targeted case review will always be made by a DWP member of staff, and we are committed to supporting all customers, especially those with vulnerabilities, throughout the review process,' the spokesperson said.
Researchers and campaigners have expressed concerns about the use of AI in welfare decisions. The Administrative Fairness Lab said previous DWP tools disproportionately flagged older people, disabled claimants and non-UK claimants, raising questions about whether historical data could introduce bias into automated systems.
Questions Grow Over AI Transparency
Researchers and campaigners argue that the main concern is not only the technology itself but also how much information claimants receive about why they have been selected for review. The DWP said some of the criteria used by the model must remain confidential to protect the integrity of the benefit system.
The Administrative Fairness Lab also questioned whether large-scale reviews could create unnecessary stress for people whose claims are ultimately found to be correct. According to the group's research, only 21 per cent of reviews identify payment errors, prompting concerns that many claimants undergo intrusive reviews unnecessarily.
Dame Sarah Healey said the trial would operate under existing safeguards and equality standards, but noted that the department would consider alternatives if the model did not perform as expected.
The DWP said AI is not intended to replace investigators. Instead, the system is designed to assist officials handling large numbers of benefit claims, with final decisions remaining with department staff. The trial has renewed debate over how government departments should balance efficiency with transparency, accountability and public confidence as artificial intelligence becomes more widely used in public services.
© Copyright IBTimes 2026. All rights reserved.


