AI, automation, and public authority

Policy Study

Artificial intelligence (AI) and automated decision-making (ADM) are increasingly used across a growing range of societal domains, including public administration. They are shaping how public authority is exercised and how civil servants carry out their work.

This policy study looks at cases from the Finnish, Norwegian and Swedish public sectors to analyse how AI and ADM work in practice. Based on interviews, desk study and cross-case comparison, the policy study suggests that the main challenges in public-sector AI and ADM are rarely technological. Instead, they are related to organisational capability, professional discretion and judgment, legal interpretation, procurement, accountability, and public trust.

Automation does not replace human judgment. Rather, across the cases, it shifts judgment and accountability between technological systems, organisations and civil servants. Automation may make routine processing more efficient, but it also creates new forms of work and organisational complexity. 

In addition to cross-case analysis, the policy study includes recommendations for public organisations and European policymakers on governing AI and ADM in ways that strengthen public value, fairness, accountability and democratic governance. 

This publication was launched at the FEPS-Nordic Digital Programme closing conference in Stockholm.