Benefit-Harm Assessment of AI in the Justice System

Abstract

This article explores the topic of the efficient use of Artificial Intelligence by judicial bodies. Beginning with the evolution from early digital tools to AI systems, it examines existing European assessment instruments, including frameworks developed by CEPEJ, the Council of Europe Committee on Artificial Intelligence’s HUDERIA methodology, and the fundamental rights impact assessment envisaged by Article 27 of the EU Artificial Intelligence Act. Current frameworks recognise that the use of AI as an element of justice system modernisation can bring benefits but also risks that require mitigation. Grounded in the principle of proportionality, the article suggests a more balanced assessment of the utility of AI linked to the justice system’s objectives. It advocates for context-sensitive evaluation and proposes a benefit-harm assessment framework to guide responsible and justified use of AI in the justice system.