Legal Argumentation and Deep Learning – Can a Robot Judge Reason?

The article aims to evaluate the possibility of automating legal argumentative practice using deep learning technology. As specific objectives, it seeks to present an overview of the legal argumentative practice and its importance for the application of the Law, through approaches linked to different legal paradigms; to discuss deep learning, its concept and characteristics, and its incursion in the application of the Law; and problematize, from the philosophical point of view of the Linguistic Turn and the ideas of Martin Heidegger and Hans Georg Gadamer, the effective performance of legal argumentation by robots, balancing its requirements with the limitations inherent to deep learning. In final considerations, the insertion, in the legal discourse, of data emerging from deep learning, is presented as a possibility of reconciling both.