Report reveals bias in DWP algorithms, calls for greater transparency

A new report published by digital rights group Big Brother Watch (BBW) indicates that the Department for Work and Pensions (DWP) is profiling millions of UK citizens annually using biased algorithms, threatening data rights while officials withhold crucial details.

The Suspicion by Design report highlights the significant expansion of artificial intelligence (AI) and algorithm-supported decision-making within the benefits system. Key findings include that approximately one million people were profiled last year by the Universal Credit Advances machine learning model, which BBW claims exhibits algorithmic bias. Internal DWP documents obtained by Big Brother Watch reveal consistent, statistically significant bias in the Advances model and other pilot tools, with disparities found for age, nationality, relationship status, and reported illnesses, which can serve as proxies for ethnicity, marital status, and disability.

Automation is also being used to select individuals for the Targeted Case Review scheme, which is expected to review over two million Universal Credit recipients by the end of the decade. Despite the scale, the BBW claims DWP refuses to be transparent regarding how individuals are flagged, with vital details about data and equalities protections either undisclosed or heavily redacted. 

BBW is calling for greater transparency on the use of high-risk data tools and demands a halt to any tool exhibiting unexplained bias.

Additional commentary by ComputerWeekly.

Training Announcement: Freevacy offers a range of independently recognised professional AI governance qualifications and AI Literacy short courses that enable specialist teams to implement robust oversight, benchmark AI governance maturity, and establish a responsible-by-design approach across the entire AI lifecycle. Find out more.