Universal Credit and Automated Decision Making: A Case of the Digital Tail Wagging the Policy Dog?

Abstract

Intended to simplify the benefit system and ’make work pay’, Universal Credit (UC) is the UK’s first ‘digital by design’ benefit. Proponents of UC highlight the greater efficiency and effectiveness of digitalisation, while critics point to costly IT write-offs and the ‘digital divide’ between people with the skills and resources to access digital technologies, and those without. Less attention has been paid to automation in UC and its effects on the people subject to these rapidly developing technologies. Findings from research exploring couples’ experiences of claiming UC suggest that automated processes for assessing entitlement and calculating payment may be creating additional administrative burdens for some claimants. Rigid design parameters built into UC’s digital architecture may also restrict options for policy reform. The article calls for a broadening of thinking and research about digitalisation in welfare systems to include questions of administrative burden and the wider effects and impacts on claimants.

Emerging modes of digitalisation in the delivery of welfare‐to‐work: Implications for street‐level discretion

“…To this extent, one of the concerns expressed by interviewees about transactional automation was that it could be used to restrict access to administrative officials via office closures and reduced opening hours. For instance, Griffiths (2021, p. 6) observes that the rollout of Universal Credit was accompanied by the closure of approximately 100 JCP offices and that ‘undermin[ing] the personal relationship between caseworkers and their clients’ can be a key driver of digitalisation reforms.…”

“…However, the nuances of how this 'new' form of engagement are changing the service encounter more broadly are yet to be fully understood and will need further interrogation. Research to date has pointed to the way digital encounters may differ from traditional public encounters (Breit et al, 2021) and to look at instruments such as the UK's online journal (Griffiths, 2021) whereas the impact of Zoom and related platforms has yet to be fully explored. While these tools were praised for increased accessibility and time-saving opportunities, they may still shape the encounter in ways that differ from concerns about automated decision-making but are no less important to understand.…”

“…In 2012, it commenced a major restructuring of working‐age benefits known as Universal Credit, which merged a series of means‐tested benefits and tax credits into one single monthly payment paid automatically. The new system introduced fully online applications to facilitate the use of algorithms and data matching to automate ‘processes for determining eligibility, assessing entitlement, calculating payment and detecting fraud’ (Griffiths, 2021, p. 2). It also incorporated some online tools, including an online journal, for jobseekers to agree claimant commitments, track their job search and compliance activities, and detail any changes in circumstance affecting their entitlement.…”

“…A deeper concern is the nebulous way ‘digitalisation’ has been used in much of the literature on the digital welfare state: as a catch‐all term for ‘all forms electronic, online, algorithmic, automated and data‐driven processes, procedures, calculations and systems’ (Griffiths, 2021, p. 13). Many qualitatively different phenomena from routine process automation to online chatbots to using artificial intelligence (AI) in predictive risk modelling are grouped together in ways that conceal what we elsewhere argue are distinct ‘varieties of digitalisation’ (Considine et al, 2022, p. 521).…”

Surveillance, Discretion and Governance in Automated Welfare

“…These calculations are usually set by a public process, and certain classes of people are automatically eligible. For example, the UK Universal Credit system determines monthly payments drawing on a range of personal circumstances, such as housing circumstances and children, but is also based on automatic sharing of data on changing income between employers and tax authorities (Griffiths, 2021). Such calculations could theoretically be done manually by staff, but the automated system enables large scale payment processing to occur regularly without applicants having to submit new evidence manually every time income changes.…”