The Right to Good Administration and Foundation Models: A European Governance Perspective and Best Practices
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
As foundation models (FMs) – widely, yet inaccurately, also referred to as Generative AI (GenAI) or General Purpose AI (GPAI) – are increasingly used in public administrations, concerns arise regarding their alignment with fundamental rights and legal principles. This paper examines how the Right to Good Administration (Article 41 of the EU Charter of Fundamental Rights) can guide the adoption of foundation models in the public sector, treating it as a broader analytical benchmark for the administrative action, as it is often reflected in domestic constitutional and administrative traditions of EU Member States. Given the black-box nature of FMs, the study focuses solely on their internal use by civil servants, excluding deployments that produce decisions directly impacting on the legal sphere of individuals. Citizen-facing applications for service delivery are likewise left outside the scope of the analysis. By linking the principles of transparency, fairness, accountability, but also privacy and cybersecurity, to the risks posed by foundation models – including opacity, bias, security vulnerabilities, and unreliable outputs – the paper assesses how public authorities can integrate these systems while safeguarding the Right to Good Administration. It combines doctrinal analysis of the EU regulatory framework, particularly the AI Act, with a comparative qualitative study of national guidelines on the internal use of FMs in Denmark, Finland, Poland, and Sweden. The paper identifies key governance requirements and offers recommendations to support rights-aligned and trustworthy internal uses of foundation models in the public sector.
Fields of Study & Subjects
Digitalization
Law
and Regulation
Administrative Law and Governance
Ethics and Social Impacts of AI
Publication details
- Journal / source
- Institutional Research Information System (Università degli Studi di Trento)
- Publication date
- 1 January 2026
- Language
- EN