Abstract

This article is a first approximation to ChatGPT’s ability to serve as a directly usable tool in the legal work of attorneys and judges. To do this, we try to determine whether that conversational chat shows patterns of reasoning in resolving conflict situations through the application of a test. It consisted of characterizing the elements with which the application built its responses against nine problems we presented to it. This analysis allowed us to determine that ChatGPT shows a pattern of preferential use of arguments of the utilitarian type, over those of a deontic nature. Furthermore, the analysis allowed us to determine the scarce use by ChatGPT of actual legal arguments. Both elements (their utilitarian orientation and the absence of legal elements) allowed us to conclude that ChatGPT is still far from being a directly usable tool in the practice of law and in the forum; however, because of the way in which it built the arguments it produced, we believe that the application shows an obvious potential to be used in functions of systematization of the Legal System.