The Impoverished Publicness of Algorithmic Decision Making.

Area of Science:

• Public Administration
• Political Science
• Law and Technology

Background:

• The integration of machine learning (ML) into public administration raises critical questions about political and legal implications.
• Digital and AI technologies are transforming democratic processes and the concept of 'the public'.
• Existing frameworks may not adequately address the challenges posed by algorithmic decision-making in governance.

Purpose of the Study:

• To explore the impact of ML on the 'publicness' of administrative decisions.
• To develop a theoretical framework for understanding public administration as communities of practice.
• To advocate for a re-evaluation of administrative law concerning algorithmic harms.

Main Methods:

• Conceptual analysis of 'publicness' within public administration.
• Examination of ML models' role in public decision-making processes.
• Theoretical development grounded in dialogical, critical, and synergetic interactions within administrative communities.

Main Results:

• ML models in public decision-making lead to an 'impoverished publicness'.
• The use of ML undermines the capacity of public administrations to function as sites of democratic construction.
• Current administrative law may not sufficiently address the detriments of algorithmic governance.

Conclusions:

• A critical reconsideration of ML in public administration is necessary to preserve democratic values.
• Public administration's 'publicness' is threatened by the uncritical adoption of ML technologies.
• Administrative law must evolve to mitigate the negative consequences of algorithmic decision-making.