Inteligência artificial e juiz-robô: os algoritmos na tomada de decisões pelo poder judiciário.

Abstract

This thesis proposes to reflect under which conditions the hermeneutics of philosophical bias, from Heidegger and Gadamer, and the Dworkin ‘s theory can serve as parameters for the use of artificial intelligence by the Brazilian Judiciary, especially in the decision-making context. It seeks to understand the (im)possibility of machines to decide lawsuits, consolidating the robot judge, replacing the human judge. The development of information and communication technologies, especially the Internet, allowed the emergence of the network society, which connects without temporal and spatial barriers. With the digital revolution, the Fourth Industrial Revolution emerged, in which artificial intelligence stands out, through the use of algorithms. This panorama impacted the Judiciary, which has been using new technologies to seek greater procedural speed, reducing the problem of judicial delays. However, alongside the numerous advantages that the use of artificial intelligence offers the Judiciary, serious risks are also revealed, especially subjectivity and the lack of algorithmic transparency, associated with the danger of non-compliance with already established rights and principles. This problem also stems from the lack of regulation of artificial intelligence in Brazil, especially regarding its use by the Judiciary. All these situations are analyzed from the phenomenological-hermeneutic approach and the use of historical, comparative and monographic procedure methods, associated with empirical research, which made it possible to consult the National Council of Justice and Brazilian Courts, through the completion of Google forms, regarding the use of artificial intelligence, especially algorithms. The 68 (sixty-eight) responses obtained were analyzed, whose practices revealed by the application of the technique of direct and systematic observation allowed the comparison between the perception of the study participants and the theoretical contribution used in the thesis. In this scenario, the opportunities and risks related to technological development challenge institutions, such as the Judiciary, which must appropriate artificial intelligence and use it, in the decision-making scenario, to support the role of the judge (and not replace it), the relevant legislative production is still necessary. The limits of the integrity of the law and of the hermeneutics of philosophical bias are revealed to be important, so that the use of artificial intelligence in the decision-making context involves, firstly, the need to guarantee the reasoning of decisions by the judge (either machine or human) to achieve the appropriate response in each process. This response also comprises an integral and coherent decision to the specific case, with the observance of the ethical and trust parameters defined by the European Union and combined with those already defined in Brazilian Law. Finally, the judicial decision must contemplate the follow-up (or supervision) of a human judge, so that the attribution of meaning to the text is produced in the context of the hermeneutic circle, with the guarantee of transparency of the entire procedure, within the fusion of horizons. meaning of the specific case, the text and the interpreter, according to the legislative initiatives addressed.