Between human judgment and algorithmic decision: the problem of noise in the context of artificial intelligence in the Judiciary

Abstract

Artificial intelligence (AI) has been progressively incorporated into the Brazilian Judiciary, especially in the context of the so-called 4th Industrial Revolution, being used by various courts with the aim of increasing the efficiency and speed of judicial services. However, the use of these technologies in the decision-making context still raises relevant debates, especially regarding the presence of noise in the judgment process and the lack of transparency of algorithmic systems. Given this scenario, this article aims to discuss, based on the work *Noise: A Flaw in Human Judgment*, by Daniel Kahneman, Olivier Sibony, and Cass R. Sunstein, how the phenomenon of noise manifests itself in both human and algorithmic decisions (mediated by algorithms), reflecting on the viability of using artificial intelligence in the Brazilian judicial decision-making context. To this end, the phenomenological-hermeneutical approach is used, considered adequate for addressing a complex issue, complemented by the perspective of Law and Literature, as well as by documentary and bibliographic research techniques. It is concluded that, although artificial intelligence can contribute to optimizing procedural management and reducing certain decisional inconsistencies, algorithmic systems do not automatically eliminate the problem of noise, and may even reproduce biases present in the data used in their operation. Thus, the use of AI in the judicial decision-making context must occur with caution, transparency, and control mechanisms that guarantee the preservation of the fundamental principles of the judicial process and the Democratic Rule of Law.