Repercussão da inteligência artificial no stf: julgamentos monocráticos de 11/05/2022 até 28/11/2024

The virtual world and the Law are becoming increasingly closer, due to the increasing emergence of several artificial intelligence (AI) tools, such as: ChatGPT, currently MARIA of the STF and previously VitórIA, with the aim of providing greater speed and effectiveness to the Judiciary. However, although such technologies provide several benefits to the performance of simple, repetitive tasks and data organization, the implementation of AI mechanisms presents several risks to the correctness and legitimacy of the system, in view of the perspective of constitutional due process and procedural transparency of monocratic decisions. Thus, we discuss the influence of artificial intelligence tools, especially on algorithms – models that abstractly represent certain real-world processes – and machine learning, to demonstrate that even tools from exact sciences, supposedly impartial, are imbued with the subjectivity of their creators and directly affected by the quality of the data provided. Therefore, it is essential to recognize the existence of algorithmic biases, given the possibility of harming the constitutive principles of the Democratic State of Law, such as due constitutional process and access to justice, so that it is then possible to think of ways to circumvent them, through machine learning and algorithmic procedural transparency.