Select committee: Public Accounts Committee
Recommendation 36
6th Report - DWP Customer Service and Accounts 2023-24
Published 31 Jan 2025
Addressed to HM Treasury
Public Accounts Committee DWP Customer Service and Accounts 2023-24 All items in this report Report on parliament.uk ↗ Inquiry on parliament.uk ↗
Recommendation (the committee's words)
DWP is implementing machine learning techniques to help it identify fraud in benefit expenditure. It has one machine learning model in operation, for new UC advance claims, alongside several others in development.69 The previous Public Accounts Committee raised concerns about the level of transparency in DWP’s use of these tools and the potential impact on claimants who are vulnerable or from protected groups.70 In December 2023, the previous Committee recommended that DWP should consider explicitly the impact of using machine learning to identify fraud on legitimate claims being delayed or reduced, the number of people affected, and whether this was affecting specific groups of people.71 DWP has since undertaken a fairness impact assessment, but has not published the results so as to conceal from fraudsters how its model operates. It says that the results did not give any cause for concern when considered alongside the safeguards in place to protect claimants.72
From: Public Accounts Committee, 6th Report - DWP Customer Service and Accounts 2023-24 (HC 354), 31 January 2025
Government response
The government responded to the whole report on 6 May 2025.
Summary written by Accountability Index from the government's response. It is not the government's own words.The government agrees with most of the Committee's recommendations in this response, but disagrees with the recommendation that it set out when it expects to meet its PIP processing target.
- On PIP claim times, the government disagrees with the recommendation, stating the department "has to balance speed against the risk of inaccuracy". It says the Health Transformation Programme will modernise PIP "in the longer-term".
- On ESA calls, the government agrees and says it is recruiting 160 colleagues by March 2025. It states that customer wait times for ESA enquiries "will be improved by September 2025".
- On counter fraud, the government agrees and says that from April 2025 the department "will begin progressive deployment of an additional 3,000 staff", expecting the full number to be in role by 2027-28.
- On machine learning, the government agrees and gives a target implementation date of summer 2025. It says the department "will support an in-confidence session" on the 2024 fairness analysis assessment.