Artificial Intelligence and LINDB: the risks of decision automation in Public Administration.

OS RISCOS DA AUTOMATIZAÇÃO DECISÓRIA NA ADMINISTRAÇÃO PÚBLICA

Abstract

The objective of this text is to investigate the impacts of technological advancement, especially Artificial Intelligence (AI), on the decision-making process of Public Administration, in light of Resolution No. 615/2021 of the National Council of Justice (CNJ). The methodology used in this article was quantitative bibliographic research, as well as qualitative analysis of decisions from both external and internal control bodies, with the aim of discussing the risks of loss of discretion and the emergence of a new cycle of decision-making paralysis in the public sector, driven by the so-called "administrative law of fear." This concept refers to the ideas of Dr. Rodrigo Valgas, who developed and refined it based on the premise of excessive oversight and the fear public managers have of being held accountable for their actions. In this context, the hypothesis is that the use of automated solutions in the management and oversight of public policies may increase legal uncertainty and reduce the legitimate decision-making space of public agents, especially in a scenario dominated by sanctioning mechanisms and strict controls. Additionally, as stated, the fear of accountability is linked to the use of technologies that, although efficient in diagnostics, are incapable of fully considering the real-world challenges faced by public managers. The Law of Introduction to the Norms of Brazilian Law (LINDB), specifically §1 of Article 22, establishes that administrative and control decisions must take into account the context, capacity, and limitations of the public agent—something that AI, by its nature, cannot fully capture.