Downloads & Free Reading Options - Results

From robo-advisors for financial investments, to self-checkout retail stores and driverless trains; automation and automated agents have become integral to various facets of daily life. In tandem with the ubiquity of automation, its importance and value in public sector contexts have become increasingly evident too. E-government platforms enhancing citizen access and engagement, and predictive analytics shaping policymaking are but a few of the myriad examples illustrating the transformative role of automation in the public sector over recent decades. As automation continues to extend its reach within governments globally, understanding the factors that influence trust in these automated systems becomes paramount. Trust, a foundational element in the relationship between governments and citizens (Levi & Stoker 2000; Hardin 2002), is crucial in ensuring the successful adoption, acceptance, and effectiveness of automated services (Bélanger & Carter 2008). This trust is not merely in relation to the functionalities of the automation, but also its governance. The importance of trust is magnified in public sector automation due to the monopolistic and unavoidable nature of the services provided (Rainey 2009). Citizens often have no alternative but to utilize public services for essential and diverse needs such as healthcare, education, and public safety. This lack of choice means that any failure in automated systems in the public domain can directly impact citizens' lives and well-being, making trust not just a preference but a necessity. The significance of trust in the public sector automation context is further amplified by the unique role of governments as service providers held to higher standards of accountability and transparency (Hood & Heald 2006). This stems from their responsibilities to safeguard public interest and welfare, requiring operations that are not only effective, but also openly scrutinized and administered. Given that the latter expectations of transparency often run counter to the black-box nature of automation (Pasquale 2015; Burrell 2016), trust acts as a cornerstone to ensure that citizens believe governments operate fairly and with accountability. However, despite the importance of trust in public sector automation, there is a dearth of empirical research (see Aoki et al. 2024, which illustrates this with a systematic literature review in the context of AI). Against this backdrop, this study advances existing conceptions of trust in automation by proposing and developing the concept of third-party vouching, a novel collective term defined as actions undertaken by individuals or entities other than potential trustors (users of the automation) or trustees (the automation or automated agent), which endorse, verify, or are perceived to provide evidence of an automation's trustworthiness; thereby facilitating trust in the automation. Some examples of what constitutes third-party vouching – and which are explicated in the later sections – include regulatory approvals by government agencies, individual users queueing to use an automated service that suggests some level of desirability of and demand for the service, and automation providers or partners associating themselves with their products and services through logo or brand placements, or media engagement. These actions that constitute third-party vouching have two main dimensions: modalities (the nature of the action) and the origins or identities of the individuals or entities taking the action. The study proceeds to test the impact of third-party vouching on trust in various hypothetical examples of public sector automation, using a novel factorial survey experiment approach fielded in Singapore. The next section outlines the efforts of existing literature to develop frameworks of trust in automation, and sets the stage for the third section to delve into the concept of third-party vouching in relation to parallel disciplines, motivated by the risk society literature; and its understated impact on user trust with ensuing hypotheses. The fourth section details the factorial survey experimental method used for empirical testing, and the fifth section presents the results in tandem with a discussion of these findings. The implications and utility of these findings within and beyond Singapore are laid out in the concluding section.