Limites éticos e constitucionais no uso da Inteligência Artificial pelos Tribunais a partir do pensamento de Hart e Dworking

• Ano de defesa: 2025
• Autor(a) principal: Lima, Felipe Carvalho de Oliveira
• Orientador(a): Chalita, Gabriel Benedito Issaac
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Pontifícia Universidade Católica de São Paulo
• Programa de Pós-Graduação: Programa de Pós-Graduação em Direito
• Departamento: Faculdade de Direito
• País: Brasil
• Palavras-chave em Português: Aristóteles; Justiça; Inteligência artificial; Decisão; Parâmetros processuais
• Palavras-chave em Inglês: Aristotle; Justice; Artificial intelligence; Decision; Procedural parameters
• Área do conhecimento CNPq: CNPQ::CIENCIAS SOCIAIS APLICADAS::DIREITO
• Link de acesso: https://repositorio.pucsp.br/jspui/handle/handle/44190

This research's main objective is to investigate whether the use of Artificial Intelligence by the country's courts complies with the principles of due legal process, respecting the constitutional and ethical limits established in our Ordinance, as well as whether its use has the purpose of improving the Judiciary. The current stage of technological development in the area of information technology allows a computer to prepare, in text form, a response to a question posed, by consulting a database guided by algorithms and systems programmed to choose the most appropriate response from among the various alternatives. Specifically in the courts, a major obstacle to Artificial Intelligence offering its potential gains lies in the segmentation of its databases – the starting point for any solution development, whether those that operate based on so-called supervised intelligence (where all parameters are offered by predefined algorithms), or those that apply deep learning, with the machine developing its own knowledge. However, it seems anachronistic to give up the potential benefits of artificial intelligence, especially when it is clear that the compartmentalization of information has deeper relationships with the preservation of spaces of power than with actual technical difficulties. This research aims to seek an eminently conceptualization of the limit to the use of techniques provided by Artificial Intelligence, based on a demonstration of its transparency and application, with a bibliographic analysis of authors who effectively interpreted its applicability and legal limits