Conference Agenda

The digital transformation of public administrations alters citizen-state interactions. Moreover, it challenges front-line workers, whose roles now involve digital interactions, presumably impacting their use of discretion. Therefore, this study explores how administrative discretion is exercised in digital bureaucratic encounters. In the extant literature, the curtailment thesis suggests ICTs reduce discretion through automation, while the enablement thesis posits that technology enhances decision-making capabilities. Research also shows mixed results on the extent of this influence and claims that digital discretion can no longer be seen as a dichotomous variable. Consequently, we analyze the following research question: How is administrative discretion exercised in digital bureaucratic encounters? Using a qualitative, multiple case study methodology, we examine discretion among front-line workers in Switzerland across different policy areas (construction permits, social benefits, and tax administration). Semi-structured interviews and document analysis provide insights into contextual factors shaping discretion. A grounded theory approach with the Gioia methodology is applied to identify emergent themes. Our findings contribute to the ongoing (dualistic) debate about the influence of the digital transformation on administrative discretion by broadening the view to the citizen-state interaction.

The ongoing digitalization of Norwegian labor and welfare services has reshaped how residents access, use, and benefit from essential public services. While digitalization offers opportunities for increased efficiency, responsiveness, and accessibility, it risks reproducing or generating new forms of inequality through digital administrative burdens. Drawing from literature on administrative burdens and digital divides scholarship, this paper explores how administrative burdens in digitalized work and welfare services are experienced and navigated by users, and how these processes are shaped by socio-digital inequalities. Our empirical analyses rest on qualitative data including: interviews with 13 street-level bureaucrats (caseworkers in the Norwegian welfare administration (Nav), advisors in other public services), and interviews with 17 young adults with migrant backgrounds who are neither in work nor in education. We find that digital administrative burdens are interconnected, but unevenly experienced across the young people. Even with access, youth struggle to understand information online and from self-service options (learning costs), while demands for digital documentation and reporting (compliance costs) generate frustration. Frustration, stress and shame (psychological costs) can lead to some abandoning services entirely, with potential negative consequences. Navigation strategies, such as using translators or ChatGPT, leveraging formal and informal supports, or channel switching can reduce but also shift burdens, and there are disparities across the youth in leveraging these strategies. Bureaucratic competence appears to be an important precondition for beneficial outcomes but is interwoven with technical skills and others individual factors. By studying these dynamics, the paper contributes to the broader literature on e-government and digital inclusion, and how digitalized welfare services could be made more inclusive for vulnerable populations.

Artificial Intelligence (AI) is emerging as one of the most significant technological innovations, with the potential to profoundly shape the design and delivery of various public services and facilitate the development of innovative service models (Mergel et al., 2024). Despite its relevance, relatively little attention has been given to public service management issues. Only recently has the use of AI in public service design and delivery become an increasingly prominent subject of scholarly debate (Madan & Ashok, 2023).

Within this context, four distinct roles of AI in the interaction between users and service providers have been identified (Osborne & Nasi, 2024). AI technologies can act as an interface, providing users with information, streamlining access, and guiding them to the appropriate services. AI can function as data processor, collecting and analyzing big data to inform better service delivery, and as a service enabler automating tasks, personalizing services, and enhancing accessibility. Furthermore, AI can serve as an active partner in the process of co-designing, co-producing and co-delivering public services.

While academic literature on AI's role as a partner in the public sector is still emerging, the topic has been more explored in the private sector (Behera et al., 2024), highlighting its potential as a promising area for future research. Exploring what role AI can have in public service delivery, how it is being used and whether it generates value contributes to advancing the literature on public administration and public service logic and may offers insights for public managers.

For this reason, we carried out a systematical literature review to synthesize existing literature across both sectors to identify the current state of conceptual and empirical studies, as well as outline a future research agenda.

To fill this gap, this study addresses the following research question:

RQ: What is the current state of research on human-AI collaboration in service delivery?

Through a systematic literature review conducted in accordance with the PRISMA framework (Liberati et al., 2009), this study examines 50 articles and 5 book chapters on the use of AI as a collaborative tool, exploring how AI tools contribute to the creation and delivery of services, and how it supports both employees and citizens. The reviewed literature is systematically charted and clustered according to several critical dimensions, including: (i) antecedents of collaboration (ii) outcomes of collaboration (iii) the role of AI in such contexts.

Based on these findings, the study maps the current state of research on this topic, identifying future research directions on AI as a tool for enhancing collaboration, and providing practical recommendations for public organizations seeking to establish more effective partnerships with AI.