Automating Universal Credit studies the automated and digital aspects of Universal Credit (UC), a social security benefit in the UK. Recipients mainly interact with UC staff through an online account, and their monthly entitlement is calculated by an automated, means-testing system that factors their personal circumstances and monthly income if they work. The study used qualitative longitudinal methods, which allowed us to understand how UC claimants experience UC in near real-time, including the systems' unexpected behaviours and errors. The study has achieved two main high-level findings. First, we found that Universal Credit creates temporal mandates by applying a fixed, monthly assessment period of earnings for working claimants. We developed the concept of temporal punitiveness to describe how this monthly period creates problems for claimants who are paid their wages on a weekly/bi-weekly basis; due to the rules of calculation within the system, this misalignment may lead to the loss of entitlement. Likewise, parents have a very narrow timeframe to submit for childcare reimbursement; if parents do not submit within certain parameters, they may not be reimbursed. Second, we argue that administrative burdens may originate at the technical layer that citizens interact with to receive a service. To evidence this argument, we identify how the mechanism for reporting earnings data to UC creates administrative burdens for working UC claimants: we found that several claimants' earnings data reported to UC for the automatic calculation was wrong and thus, claimants needed to start a dispute process with the DWP. A FOI request we submitted to the DWP revealed that this happens to 5-6% of claimants each year. UC’s automated tools function at times as gatekeepers to social security entitlements, raising concerns about the current design of UC as a form of social security.In the UK, the Department for Work and Pensions (DWP) uses automated systems for Universal Credit (UC), the country's largest social security payment, to determine eligibility, calculate monthly benefits and detect fraud. Unique to UC, automation determines monthly pay based on a complex set of means-testing variables and data linkages with other departments. What are the policy rationales behind UC’s system? How do claimants experience the monthly means-tested payment? And how does the system shape the way DWP carries out its welfare policies? This project contributes to our understandings of digital welfare broadly, exploring how claimants experience these systems and can inform their technical design.

During the first six months of the study we interviewed 18 members of staff at charities and civil society activists across the UK. The remainder of the study involved Scottish claimants who were in-work and took part in qualitative longitudinal research (QLLR) to understand the experiences of Universal Credit recipients interacting with the automated aspects of this benefit for six months to a year. 25 claimants took part; we carried out entry, mid-way, and exit interviews (47 total) combined with prompted bi-weekly text message updates from participants on their interactions with UC; many would send back text responses and screen shots of the UC account, both of which we captured as field notes (145 pages total). We also carried out one-off interviews with 27 claimants who did not qualify for the longitudinal study because they were not required to work to qualify for UC. We recruited all participants through local Scottish charities.