Justiça preditiva e proteção de grupos vulneráveis: análise dos Vieses discriminatórios decorrentes da automação judicial

• Ano de defesa: 2025
• Autor(a) principal: Cruz, Ramon Aranha da
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Ciências Jurídicas; Programa de Pós-Graduação em Ciências Jurídicas; UFPB
• Palavras-chave em Português: Direitos humanos; Justiça preditiva; Grupos vulneráveis - Discriminação; Human rights; Predictive justice; Vulnerable groups; Discrimination; CNPQ::CIENCIAS SOCIAIS APLICADAS::DIREITO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/36742

This doctoral investigation focuses on the new forms of vulnerability experienced by socially vulnerable groups through the use of artificial intelligence in decision automation. Based on this problem, the hypothesis posits that the development of algorithms originally based on Human Rights will directly lead to the end of discriminatory acts resulting from judicial automation. The general objective of this thesis is to analyze how it would be possible to implement tools to protect vulnerable groups subjected to assessment by artificial intelligence software in Criminal Justice. The specific objectives are: a) examine the process of creating artificial intelligence tools used as aids in state judicial activity and predictive justice to understand their functioning; b) study the ethical relevance of using intelligent software in decision- making processes by public and private entities; c) analyze legislative texts that aim to regulate the matter and whether the response to such violations can be obtained through regulation; d) investigate the process of vulnerability of social minorities, their causes, and main factors of marginalization; e) assess and propose public policies to prevent the victimization of vulnerable groups subjected to automated scrutiny by the Judiciary and other public entities. From a methodological perspective, the thesis is classified as exploratory and follows a hypothetical-deductive method, using bibliographic and documentary research as a technique, through books, scientific articles, and legislation. At the end of the research, it was possible to confirm the proposed hypothesis, both through the scientific evidence demonstrated and through the mass adoption of the regulatory system by large States, highlighting the need for regulation to protect vulnerable groups subjected to automated decisions in the Judiciary.