Costa Filho, Bento Alves da2026-07-132026-07-132025CAVALCANTE, Nilvânia Maria de Fátima. Uso da inteligência artificial no poder judiciário: uma análise sob a perspectiva dos integrantes do setor processual em 2ª instância do Tribunal de Justiça do Estado de Goiás. Goiânia (GO), 2025. 83 f. Dissertação (Mestrado Profissional em Administração) - Centro Universitário Alves Faria, 2025.http://repositorio.unialfa.com.br/handle/123456789/803This study is situated within the contemporary debate on the digital transformation of the public sector, with a specific focus on the adoption of artificial intelligence (AI) tools by the Brazilian Judiciary. Its general objective is to investigate the perception of the Court of Justice of the State of Goiás (TJGO) regarding the use of artificial intelligence in adjudicative activity at the appellate level. More specifically, it aims to identify the main technical and legal challenges faced by professionals, demonstrate the level of acceptance of AI tools in light of the guidelines of the National Council of Justice (CNJ), and verify whether the use of AI has positively contributed to the effectiveness, efficiency, and overall performance of judicial work. The research addresses a relevant gap in the literature, since, although there are studies describing AI tools implemented in Brazilian courts or discussing their normative and theoretical aspects, empirical investigations analyzing the perceptions of professionals who actually operate these technologies in daily judicial practice remain scarce, especially at the appellate level, a context characterized by particular features related to case volume, composition of adjudicative bodies, and work dynamics. Methodologically, the study adopted a qualitative, exploratory, and descriptive approach. Data were collected through semi-structured interviews with 21 clerks and assistants to appellate judges, organized into four thematic blocks: participants’ profile; relationship with, acceptance of, and adaptation to AI; technical and legal challenges; and impacts on productivity, effectiveness, and efficiency. Content analysis (Bardin) was applied to the corpus, using open and axial coding guided by both a priori and emerging categories, with the Technology-Organization-Environment (TOE) framework serving as the analytical lens. The main findings indicate that the adoption of AI tools at the TJGO, particularly AGAIA and Berna, developed internally, follows a predominant pattern of initial resistance followed by progressive adaptation, with largely positive acceptance among professionals. All 21 participants unanimously stated that AI complements, rather than replaces, human judicial work, emphasizing that ethical and legal responsibility for decision-making remains inherently with the judge. Identified challenges include the occurrence of hallucinations, instability and operational slowness during peak hours, insufficient coverage of criminal law compared to civil law, difficulties in crafting effective prompts, and concerns regarding data protection in cases under judicial secrecy. An asymmetry was also identified between formal knowledge of CNJ resolutions and their effective incorporation into daily practice, with a significant portion of interviewees being unaware of or having only superficial knowledge of CNJ Resolution No. 615/2025. Regarding productivity impacts, although the data indicate clear and unequivocal gains in terms of efficiency, effectiveness proved to be conditioned by the level of critical engagement of the professional, as AI substantially contributes to the textual and linguistic quality of documents, but its contribution to substantive legal analysis is limited: the tools do not engage in legal reasoning, do not assess facts, nor individualize conduct. A structural divide observed throughout the research shows that benefits are significantly more pronounced in the civil area, where normative matters and repetitive demands prevail, than in the criminal area, which requires individualized analysis of facts, evidence, and personal circumstances. The study also identified, as an emerging category, the risk of professional burnout as an unanticipated side effect of increased productivity, suggesting that expanding productive capacity without proper management of cognitive demands may compromise the sustainability of the observed gains.ptInteligência Artificial (IA)Poder judiciárioTransformação DigitalEficácia jurisdicional