Responsible artificial intelligence and Justice 5.0 in CNJ Resolution No. 615/2025

Abstract

The promise that artificial intelligence (AI) will render justice faster, more efficient, and more personalized coexists with the risk that algorithms may intensify bias, opacity, and inequality. In the Brazilian Judiciary, the incorporation of predictive and generative systems typical of the Justice 5.0 paradigm makes it urgent to define safeguards capable of preventing adverse impacts on fundamental rights, democratic legitimacy, and institutional sustainability. In this context — marked by the absence of a general national AI law and by the legislative deliberation on Bill No. 2,338/2023 — this article examines the extent to which CNJ Resolution No. 615/2025 constitutes a sectoral regulatory framework for responsible AI in the Judiciary. The study adopts a qualitative, deductive approach, combining a normative–documentary analysis of the Resolution with an integrative review of recent literature (2018–2025) on algorithmic risks, risk-based governance, and justice, articulated through interpretive triangulation. The results indicate that the norm consolidates a systemic and risk-proportionate regulatory architecture, including explicit prohibitions, requirements for meaningful human oversight, mechanisms for transparency, traceability, and auditability, personal data protection measures, and institutional arrangements for plural oversight. The article concludes that the Resolution contributes to aligning the Brazilian Judiciary with the international paradigm of trustworthy AI and provides operational parameters for future national AI regulation and for the design of high-impact public policies in the justice ecosystem.