AI and the quest for legitimacy in the modern society How does the use of artificial intelligence in public governance impact institutional legitimacy? The case of Norway.

Abstract

This study examines the impact of using Artificial Intelligence (AI) in public governance on institutional legitimacy by using a multilevel mixed methods research design. Institutional legitimacy is an essential ingredient for organizational success and survival. However, as our societies become more complex, more advanced solutions are sought to address the challenges that emerge. Consequently, there is a growing desire to incorporate AI into public governance to mitigate these challenges. Therefore, this study aims to research the potential impact of using AI in public governance through the research question: How does the use of artificial intelligence in public governance impact institutional legitimacy?

Institutional legitimacy in this thesis is understood and studied as a multilevel and multifaceted concept. Hence, central is the reflection on the following question: Legitimate to whom? Approaching legitimacy as such demands the integration of several levels of analysis to produce an adequate answer. In this thesis, I explore the political, organizational, and individual levels. Moreover, when discussing the potential challenges the incorporation of AI in public governance may generate, one of the most controversial cases remains unexplored: high-discretion tasks. Thus, to answer the research question, I combine a survey experiment from the Norwegian Citizen Panel at the individual level with discourse analyses of public documents at the organizational and systemic levels.

The findings indicate that there is a schism in the understanding and legitimacy of using AI in public governance. Political leaders frame AI as a necessary innovation, a “magic concept” necessary for the country’s future, and the regulations needed as a “necessary evil”. Meanwhile, the citizens consider the regulations an essential part of the process. In the “eyes” of the public, humans remain the most legitimate decision-makers, but using AI in high-discretion tasks may be legitimized through the use of accountability and transparency measures. Finally, organizations find themselves in the middle, realizing that AI has immense potential while also acknowledging the dangers associated with the deployment of AI in public governance.