Impact of Digitalization on discretionary decisions
Underlying Problem
The work of public administrations is increasingly supported by digitalized processes and services and there is a political will to obtain even further benefits. Digitization yields ambiguous consequences. On the one hand, digitally supported and automated public services possess great advantages including efficiency, improved use of experience and absolute objectivity. The increase in efficiency can be found in the execution of processes, as information is processed automatically, improvement in use of experience, since all information continues to be available for any subsequent decision, and absolute objectivity, as far as unambiguous factual situations are concerned (Bullock et al., 2022; Young et al., 2019). On the other hand, AI can reproduce malignant stereotypes in individual applications (Hundt et al., 2022), has difficulties to apply public service ethics (Leslie, 2019) as well as posing a potential existential threat (Bullock, 2019). Especially decisions in public administration, so-called administrative acts, are supported by digital provision of information, virtual interactions and automated information processing (Busch, 2018; Busch & Eikebrokk, 2019). These administrative acts are deterministic to varying degrees, which means that public servants are given discretion to decide on administrative acts after further examination and taking additional information into account. The digital processing and monitoring of administrative acts provides precise information and assistance for processing, which can facilitate decision-making or even make recommendations for decisions (Busch, 2018). Taking into account the weaknesses mentioned above, digitization and the use of AI in administrative acts must be considered carefully (Bullock et al., 2022; Young et al., 2021).
Research Question
Starting points for this research are characteristics of tasks, that provide comparative advantages for either human or AI (Bullock, 2019), features of public services that explain the acceptance of digital discretion by public servants (Busch & Eikebrokk, 2019), and the diagnosis, that application of AI and digitized processes is not favourable in every case Young et al. (2021). Against this background, I intend to investigate, how human and technological decision-making skills can be combined to obtain good discretionary decisions. Therefore, I will test underlying mechanisms and moderating conditions.
Aims and benefits
This research aims to provide public administrations with insights to improve administrative performance while maintaining the level of human labour input. In addition, these insights ensure the maintenance of quality in administrative services and the prevention of bad administration by digitization. On an individual level, I aim at enabling a human centered design of digital work conditions in public administration. Including both perspectives of public servants and citizens/clients.
Relevant theories
Public servants’ perceptions of digital tools, their application and automation range from welcome support and relief in complex situations to the feelings of disenfranchisement and obstruction of autonomous decision-making (Busch, 2018). In order to capture the perception of different forms of digitization and the use of AI by public servants, the Jobs-Demands-Resource-theory (Demerouti et al., 2001) will be used as explanatory model. This widely researched and integrative theory describes how characteristics of the work environment can be experienced as resources or demands that shape work experiences and outcomes and has been successfully applied to explain ambiguous digitalization effects(Demerouti, 2020). Whether digitalization is perceived as beneficial or hindering largely depends on
public servants’ acceptance of these technologies which can be explained by the Unified Theory of Acceptance and Use of Technology (Venkatesh et al., 2003). This theory specifically helps to further investigate the motivators for compliance with digitalized processes among public servants.
Theoretical/analytical approaches
Central constructs of these research questions (see Figure 1 [in PDF]) can be operationalized building on well-researched theories. In order to obtain a measure for the different forms of digitization of administrative procedures, the five dimensions model of digital work Kaesmayr et al. (2021) is applied which differentiates information, interaction, integration, data processing, and transaction. Differentiation of discretionary decisions can serve as a starting point on predicting the superiority of
human or digital discretion and pointing to an integration potential. Administrative procedures and decision-making whether digital or not depend on public servants’ individual actions, organizational structures and processes and national legislation. Hence, the complex interplay of multiple levels of administration according to Gofen et al. (2019) and Moynihan (2018) need to be taken into account. My primary focus will be the street-level-bureaucrats. To understand the quality of discretionary
decisions multiple stakeholders need to be considered.
Research design
The research design follows a mixed-methods approach (Creswell & Creswell, 2018). First, influences on discretionary decisions are collected in a qualitative study by conducting interviews with public servants. The collected data will be used for typifying and describing discretionary decisions and specific challenges in different forms of digitization such as automated provision of information on individual cases, optional support with relevant background information, or semi-automation of discretionary decisions. Subsequently, building on survey data, regression analyses and structural equation modelling will be used to test the proposed and identified predictors, moderators, and mediators of decision quality. The evaluated influences are examined and quantified via an online survey. Finally, experiments will be used to test different digital solutions or uncover specific underlying cognitive or motivational mechanisms.
Expected results
I expect an ambiguous impact of digitalization on quality of discretionary decision and that beneficial effects depend on individuals’ competencies and technology perceptions, forms of digital support/digitalization and characteristics of the decision, and potentially on the rater. Expected results include empirical validation and substantiation of findings by Lipsky (2010) and Bullock (2019)
Problems and challenges
Although practitioners or politicians might wish for easy answers, the complexity of the topic will lead to various and differentiated implications. Varying degrees of e-Government maturity will result in different influences of digitalization on discretionary decisions. To be able to draw comparisons between administrations, the effect must be controlled. Further challenges arise from the definition of quality in administrative action, as stakeholders at different levels of government have different
standards for good decisions