A default framework for fundamental rights impact assessment

Abstract

The Fundamental Rights Impact Assessment (FRIA) mandated by the EU Artificial Intelligence Act (AIA) has emerged as a critical tool for evaluating the normative implications of introducing certain high-risk AI systems into the EU market. However, the AIA does not prescribe any specific methodology for conducting FRIAs. To address this gap, this paper proposes a default framework based on defeasible normative reasoning that aims to provide practical guidance in determining whether an AI system falls within an acceptable threshold for fundamental rights protection. By elaborating on how the default framework can serve as a foundation for an appropriate methodology, this study introduces a novel qualitative approach to legal risk assessment.