A justiça como harmonia na inteligência artificial: uma leitura platônica aplicada à ética tecnológica

Abstract

The present work situates itself at the interface between technological ethics and political philosophy, investigating how Plato’s tripartite soul theory can underpin the assessment of justice in Artificial Intelligence systems. By reviewing contemporary AI challenges—such as algorithmic bias, lack of transparency, and ethical oversight—a need was identified for criteria grounded in philosophical principles. Drawing on The Republic, this study proposes a model of an Artificial Soul composed of three functional modules: a rational component (Director AI) for truth-seeking and decision guidance; an irascible component (Defender AI) to ensure justice criteria; and an appetitive component (Producer AI) for action execution. The central hypothesis is that this analogy defines justice as internal harmony among the modules, offering an alternative to both utilitarian and Kantian approaches. Based on this model, the work proceeds to propose a framework that includes defining functional balance metrics and comparing them with existing ethical models, revealing that injustices can be seen as structural disharmonies to be corrected. In a broader context, the research contributes to ethical AI governance discussions and suggests pathways for empirically driven implementation and future dialectical investigations.