JUDGE HERCULES AND ARTIFICIAL INTELLIGENCE: STRUCTURAL LIMITS OF DECISION-MAKING AUTOMATION IN THE BRAZILIAN JUDICIARY

Abstract

The implementation of artificial intelligence systems in the Brazilian Judiciary raises questions about the limits of automation in adjudication. This article investigates whether judicial decisions can be replaced by algorithms, using the distinction between Ronald Dworkin's interpretive model and H.L.A. Hart's positivism as a theoretical framework. It is demonstrated that while AI can optimize informational tasks, it cannot perform the value judgments that characterize adjudication in Dworkin's view. Even under Hart's perspective—which admits mechanical application in the "core of certainty"—automation faces limits in penumbral zones, where judicial discretion requires legitimacy that opaque algorithms lack. It is concluded that technology must be understood as an auxiliary tool, not a substitute for human interpretive processes, especially in cases involving fundamental rights.