Between Algorithms and Guarantees: A Critical Review of Artificial Intelligence in Criminal Proceedings

Abstract

This review article offers a critical and multidisciplinary perspective on the use of artificial intelligence (AI) in criminal proceedings, based on a central concern: under what conditions can this technology be integrated without compromising procedural safeguards or the principle of legal truth? To address this question, the analysis focuses on three key dimensions: (i) the ethical principles that should guide the design and implementation of AI systems, with an emphasis on explainability and accountability; (ii) the epistemic justification criteria required to validate the knowledge produced by these systems in judicial settings; and (iii) the specific risks associated with their application in predictive decisions that may affect fundamental rights. Based on this analysis, the article identifies the limits and possibilities of a “procedurally trustworthy” AI, as well as the normative, ethical, and epistemic conditions that should guide its integration into contemporary criminal justice systems.