Inteligência artificial centrada no ser humano e sua aplicação no poder judiciário brasileiro: o exemplo do projeto do TRT da 4ª Região

Abstract

Artificial Intelligence (AI) is a reality in the judiciary, especially due to the advantages in efficiency gains. However, its use presents risks, such as the replacement of human decisions by the machine, biased results, lack of transparency and explainability of the algorithms, among others. In this context, the research deals with AI in judicial activity from the perspective of human centrality, a guiding principle for the development and advancement of technology. Empirical observation, based on personal experience in academia and professional activity with the Labor Court of the 4th Region (TRT4), which has an AI system in use, reveals that legal practitioners have a mere glimpse of AI in the judiciary field. The question, then, is: from the point of view of human-centered AI, how to add greater knowledge and reliability to the system used in TRT4? As a research hypothesis, textual content is proposed to support booklets to be addressed to TRT4, containing information about the adopted AI, intended for the external (lawyers, jurisdiction, society) and internal (servants, managers, magistrates) audiences. The general objective of the study is to explore the application of AI in judicial activity from the perspective of human centrality, considering that transparency, by providing information to the community, adds greater reliability and supports the idea of a human-centered AI. The specific objectives, which will compose the three chapters of the work, are summarized as follows: a) in the first chapter, to present concepts related to AI and the principle of centrality in the human being; b) in the second, to make considerations about the digital process, notably the electronic one, with emphasis on the Labor Court, and c) analyze uses and perspectives of AI in judicial activity, including hypotheses of efficiency gains and mapping risks; d) in the third chapter, explore data from the AI system that integrates the Management Shock Project in the Magazine Resource, used in TRT4, and e) submit textual content to TRT4 to, if it so wishes, publicize data about the AI system in use to the legal community. In the development of the research, a qualitative approach is adopted, with a deductive approach, using an exploratory method in bibliography, documents and regulations; as well as qualitative and empirical methodology in a case study, considered secondary to the bibliographic and documentary method, adopting the Management Shock Project in the TRT4 Magazine Resource for this purpose. It is concluded that AI in the judiciary has numerous advantages, but also risks, which can be addressed from the beginning of human centrality, with the adoption of audits, accountability and good practices. The proposal to prepare a booklet for the recipients of AI systems in the judicial field is in line with the idea that the spread of knowledge about AI, its use, purposes and results translate transparency and add reliability to the system and institutions, integrating the conception of AI centered on the human person, aiming at the promotion of fundamental rights, democratic values, and the well-being of the community.