Literacy, judicial justice and artificial intelligence

Abstract

The study identifies various existing difficulties in effectively implementing artificial intelligence (AI) techniques in judicial processes. The approval of regulations, such as Royal Decree-Law 6/2023 in Spain, or the implementation of projects like Assis in Brazil, prove insufficient for judges and legal professionals to effectively utilize these technologies. Several reasons are highlighted for these challenges. (1) Judicial proceedings and sentences must be approved by all participants, including lawyers, prosecutors, and judges. The focus should be on the specific conflict and relevant legal texts, not on AI-generated models based on past data, which may be biased. Additionally, (2) because AI technologies are not designed to assist in the specific tasks required by judges and other officials responsible for processing judicial proceedings. Judicial processes are governed by strict constitutional, procedural, and substantive norms that AI systems are not prepared to handle without significant human supervision. (3) The study also refers to critical experiences in other countries and to the opinions of the General Council of the Judiciary in Spain, which point out the lack of precision in Spanish regulation (Royal Decree-Law 6/2023) regarding the use of AI in judicial activities. The study concludes that promoting AI technologies in the elaboration of judicial sentences therefore requires, not only regulatory approval, but also comprehensive reforms of existing norms, the creation of precise complementary regulations, and even the implementation of permanent adequate literacy campaigns for technicians, politicians and jurists.