Automated decision-making in the public sector. Artificial Intelligence vs Administrative Intelligence?

Inhaltsverzeichnis (Table of Contents)

• Introduction
• State of research
• Agency theory
  • Theoretical context
  • Principal-agent problems
  • Control methods
  • Agency theory with artificial agents
    • Definitions and agency of algorithms
    • Principal-agent problem with artificial agents
    • Control methods
• Agency theory in political sciences
  • Case-study: France
    • Public institutions in France
    • Control methods
    • Algorithms in the French public sector
  • Changes in agency when ADM is used in the public sector
    • Potential risks of using artificial agents
    • Responsibility of decisions taken with the help of ADM
    • Consequences on the organisation of administrations
  • Changes in control methods when ADM is used in the public sector
    • Selection and development of the agent
    • Supervision and monitoring of the agent
• Conclusion

Zielsetzung und Themenschwerpunkte (Objectives and Key Themes)

This paper investigates the impact of automated decision-making (ADM) on the governance and responsibility of public administrations. Specifically, it explores whether the introduction of ADM transforms the agency of administrations, and if so, why and how this change impacts the control methods required to supervise their actions. The paper uses agency theory to analyze this relationship, focusing on the delegation of decision-making between public institutions, private companies, and citizens. It examines the principal-agent problem that arises when these entities have conflicting interests and the principal cannot directly ensure that the agent is acting in their interest.

• The role of automated decision-making (ADM) in public administration.
• The implications of ADM for the agency of public administrations.
• The application of agency theory to analyze the relationship between public administrations and artificial agents.
• The impact of ADM on control methods used to supervise public administrations.
• Case-study analysis of France's approach to ADM in the public sector.

Zusammenfassung der Kapitel (Chapter Summaries)

• Introduction: This chapter introduces the topic of automated decision-making (ADM) in public administrations, using the example of predictive policing in Marseille, France. It discusses the potential benefits and risks of ADM, highlighting the importance of addressing the governance and responsibility implications of this technology. The chapter outlines the research questions and the chosen approach, which involves applying agency theory to the relationship between public administrations and artificial agents.
• State of research: This chapter reviews existing literature on the topic of ADM in public administrations, providing a context for the current study. It explores various aspects of ADM, including its theoretical foundations, practical applications, and ethical considerations. The chapter also discusses the limitations of current research and identifies potential areas for further investigation.
• Agency theory: This chapter introduces the concept of agency theory, outlining its key elements and applications. It focuses on the relationship between a principal and an agent, exploring the potential for conflicts of interest and the need for control methods. The chapter discusses the application of agency theory in political sciences and examines its relevance to the study of ADM in public administrations.
• Agency theory with artificial agents: This chapter delves into the specific application of agency theory to the context of artificial agents. It defines the agency of algorithms and discusses the principal-agent problem when artificial agents are involved. The chapter examines the control methods that can be employed to ensure the alignment of values and interests between principals and artificial agents.
• Agency theory in political sciences: This chapter explores the relevance of agency theory to political sciences, providing a framework for understanding the relationship between public institutions, private companies, and citizens. It examines how agency theory can be applied to analyze the delegation of decision-making in the public sector and the potential for conflicts of interest between different stakeholders.
• Case-study: France: This chapter focuses on France as a case study, examining the country's approach to ADM in the public sector. It describes the relevant public institutions, control methods, and algorithms used in various domains. The chapter also analyzes the impact of ADM on the agency of French public administrations and the control methods employed to ensure accountability.

Schlüsselwörter (Keywords)

This paper focuses on the intersection of automated decision-making (ADM), public administration, agency theory, artificial intelligence (AI), and control methods. It examines the implications of ADM for the governance and responsibility of public administrations, specifically focusing on the principal-agent problem when artificial agents are involved in decision-making processes.