Decision-making automation and the justice problem in Brazil

Abstract

This article explores, through a qualitative and exploratory bibliographic review, the tensions and challenges brought by the automation of decision-making processes within the Brazilian legal system, with a particular focus on the judiciary. It analyzes how automation, particularly through artificial intelligence, reshapes traditional notions of justice and legal legitimacy. By proposing a dichotomous framework between justice-as-legal-positivism and justice-as-social-value, the authors investigate the consequences of reducing legal reasoning to computational logic. The study revisits the emergence of the so-called “postmodern justice,” which arose in response to mass litigation and the pressures of the digital economy, and which led to technocratic and efficiency-oriented responses such as predictive systems and automated legal classification tools. The paper also highlights the conceptual and operational issues of algorithmic opacity ("black boxes") and the semantic reduction of legal norms into machine-readable code, calling into question the interpretability, fairness, and explainability of such systems. Drawing on authors such as Birhane, Danaher, and Morozov, the article warns against a creeping “algocracy” and the bureaucratic legitimation of non-democratic decision-making processes. Finally, it underscores the risks of uncontrolled adoption of AI systems by both public and private actors and argues for a robust legal framework to ensure transparency, participation, and substantive justice in the use of automated decision systems.