Machine learning and the maximum support to judges

Abstract

This article is written for jurists and it brings at least one relevant indication to the technologists. E-contemporaneity will be distinguished by learning algorithms (apprentices). They can already fully support the judge if they are setted to learn as second-order observers (Luhmann). For each judge, an apprentice is the proposal that respects the brazilian constitucional procedural model. The bibliographic research proves the presence of the algorithms as true actants (Latour, Law, Callon) in legal scenarios. And it unveils the academic (Teubner) and jurisprudential debates on the ways of dealing with such technological entities. The judicial procedure in Brazil has always incorporated technical innovations (writing, typing, computers, internet). Now it must adopt the apprentices, experts in text mining, to maximize the support to judges (focus of this work), as well to other law operators. This article also suggests a revisit to the Theory of the Subject and to the theory of Procedural Acts. It is hoped to contribute to: (1) the perception and understanding of the algorithms evolutionary phenomenon - qualitative and paradigmatic; (2) facilitate the acceptance of their presence in legal scenarios and (3) foster readiness and studies to incorporate them adequately in theory and practice of the process. The method used is  the inductive one, but with predictive hypothetico-deductive freedoms.