AI in public administration: Relationships between algorithmic regulation and decision automation in the Danish AI “signature projects”

Resumé

In the period 2020-2022, the Government, Danish Municipalities (KL), and Danish Regions invested in 40 "signature projects" aimed at gaining experience with artificial intelligence (AI) in the public sector. This article utilizes an expanded typology from the framework of "algorithmic regulation" to investigate what forms of algorithmic regulation the signature projects represent and how AI is developed to support and/or automate decisions in the Danish public administration. The analysis contributes conceptually and empirically to the existing literature by (a) developing an expanded typology to classify AI as regulation and decision automation and (b) demonstrating the connection between regulatory and automated decision forms across all Danish AI signature projects. This opens a nuanced discussion of how AI promotes different forms of algorithmic regulation and decision automation in various administrative areas, and the implications of this for the relationship between the state and the citizen.