What is the effect of AI technology usage by bureaucrats under problematic conditions on the perceived legitimacy of bureaucratic decisions? Scholars argue that AI usage potentially exacerbates the...Show moreWhat is the effect of AI technology usage by bureaucrats under problematic conditions on the perceived legitimacy of bureaucratic decisions? Scholars argue that AI usage potentially exacerbates the negative consequences of misused bureaucratic discretion. Others suggest that AI usage might curtail bureaucratic discretion and increase outcomes of equity and efficiency. Existing empirical research demonstrates no significant difference between the perceived legitimacy of AI-assisted and human decision-making. This study aims to determine the effect of AI usage on the perceived legitimacy of bureaucratic decisions made under problematic principal-agent dilemma conditions. This effect is assessed across 96 survey respondents from the University of Leiden and the University of Amsterdam using experimental factorial survey analysis. The results of this research indicate that AI usage in decision-making has a significant positive effect on perceived legitimacy (p < 0.001). The main implication of this research is that AI usage can plausibly alleviate the impact of consequential bureaucratic decisions on perceived legitimacy by obscuring bureaucratic discretion. A second implication is that AI usage in bureaucratic decision-making exerts a notable effect on the perceived efficiency of bureaucratic decisions.Show less
This study uses the Technology Enactment Framework (TEF) to explore how the Customs administrations of Türkiye and the Netherlands design and implement AI technology in their automated detection...Show moreThis study uses the Technology Enactment Framework (TEF) to explore how the Customs administrations of Türkiye and the Netherlands design and implement AI technology in their automated detection pilot projects to facilitate legal trade and detect illicit trade at border crossing points. It also displays what organizational elements can be learnt from the two cases to add to the limited literature in AI applications in the public sector. At the end of the research, AI Technology Enactment Flowcharts for the two administrations, a Comparison Table and a Proposed Extended AI TEF for Customs were developed. Validating the TEF, the research also identified AI-technology specific organizational forms, institutional arrangements and project processes facilitating or impeding the pilot projects in the two customs administrations.Show less
Since the rise of artificial intelligence (AI) and its increased use in the public sector, there has been a two-fold debate in practice and academia about the relationship between transparency and...Show moreSince the rise of artificial intelligence (AI) and its increased use in the public sector, there has been a two-fold debate in practice and academia about the relationship between transparency and trust. On the one hand, transparency can create an open government culture and improve trust; on the other hand, transparency confuses citizens and leads to delegitimization of government. This research contributes to the debate by examining the effect of algorithmic decision-making transparency on institutional trust and procedural fairness's role in this relationship through a quantitative vignette study focusing on enforcing parking fines. The study distinguishes itself from others by measuring institutional trust by three dimensions (competence, benevolence, and honesty) and transparency by two dimensions (accessibility and explainability). The results reveal that: (1) accessibility increases institutional trust and trust in the competence and benevolence of the institution; (2) explainability increases institutional trust and trust in the competence of the institution; (3) procedural fairness negatively affects the relationship between accessibility and institutional trust and between accessibility and trust in the competence of the institution; (3) procedural fairness positively affects the relationship between explainability and trust in the benevolence of the institution. These results imply that institutional trust is multidimensional and can be influenced differently but that access and explainability of decision-making increase trust. Furthermore, local governments can make greater use of the Dutch algorithm registry to provide citizens with access and explanations of decision-making to enhance trust in the institution.Show less
