Artificial intelligence and precautionary guardianship. Special reference to pre-trial detention

Abstract

This research focuses on the study of the use of AI in pretrial detention decision making. Specifically, it analyzes the way in which judges usually assess the requirements of pre-trial detention and questions whether AI systems can replicate such assessment processes, improving the efficiency of decision making. To this end, we analyze the biases that judges incur when deciding on pre-trial detention, assessing the purposes and dangers that justify it, and how AI tools could replicate and, when possible, improve such decisions. Finally, the paper reflects on the risks that the introduction of AI systems could pose to procedural rights and guarantees, trying to establish the precise precautions or safeguards that should flank the advance of AI in criminal justice to achieve an incorporation of science in justice that is respectful of the fundamental rights of the defendant. In short, this paper hinges on two key questions to be answered: 1st) Can the use of AI serve to improve the efficiency of the decision-making process regarding the adoption of pre-trial detention? 2nd) What risks does the use of AI in this context pose to the rights and guarantees of the defendant and what precautions should be taken to prevent such risks?