Department for Work and Pensions (DWP) AI Systems Allegedly Discriminate Against Single Mothers

The Universal Credit system, managed by the UK Department for Work and Pensions (DWP), serves 6.7 million claimants as of April 2024 and has operated as a 'digital by default' system since its 2013 launch. The system uses automated processes to determine eligibility, entitlement, and detect fraud. A Freedom of Information request from the University of Edinburgh revealed that nearly half of working Universal Credit claimants are single parents, with just under 90% being women. Researchers identified three main problems with the automated processes: mistakes from invalid earnings information, delayed childcare reimbursement causing hardship, and inconsistency in calculation dates and paydays. The algorithm fails to properly account for payment frequency, leading to over or underestimation of earnings and causing financial turbulence for low-income recipients. A University of Bath study found that over half of Universal Credit households experienced payment variations of £400 or more monthly, with a quarter seeing £600 variations. Since single mothers represent around 40% of working claimants, they are disproportionately affected by these algorithmic flaws. Researchers noted that algorithmic discrimination is difficult to challenge due to government lack of transparency, with only 8 algorithms officially recorded despite an estimated 40 being used across government departments.