Gestão de riscos no despacho aduaneiro de importação: inteligência artificial como instrumento e agente de controle

Abstract

Administrative Customs Law is the subsystem of Administrative Law that must regulate, through formal law, the functional exercise of customs administration and that must protect not exactly its mechanisms, but the purposes of the Democratic Rule of Law. Thus, the notion of customs control as a classic police power is naturally led back to the idea of social and economic ordering activity, which not only restricts, but shapes individual rights, in accordance with the possibilities and limits established in the Constitution. Among the diversity of customs controls and the administrative powers that presuppose them, customs clearance is the most prominent, known as the primordial, cornerstone and basis institute of Customs Law, since without it it would become impossible to announce the consummation of its object: the international movement of goods. As a typical administrative procedure comprised of a series of acts or stages, the order must be subject to the control of its conformation and the legal dictates to which it is subject, especially as it establishes legal relationships based on the binomial power and subjection. Based on the hypotheticaldeductive method, the general objective of this work was to research the application of artificial intelligence in the customs import clearance procedure, with an epistemological approach to the stage of selecting customs declarations based on the risk-based operational approach. Despite the fact that customs selection is supported by a sophisticated and confidential risk management process, digital technologies do not take away its intrinsic characteristic of an administrative act. Automated selection decision-making processes, which are discretionary, by their nature, must have efficient legal mechanisms that allow their control, especially in cases where their use implies a violation of principles such as equality, legality, transparency and due process, constitutionally enshrined values. It is concluded, therefore, that the combination of these new paradigms – risk-based management and the use of artificial intelligence – poses risks and calls for at least three regulatory plans: the first, of a general nature, consisting of regulating the use of AI in Law, a topic still subject to definition by the Brazilian Legislative Branch. The second and third, of a more specific nature, concern the institution of formal law that defines customs risk management criteria, but which mainly determines the forms of maintenance, supervision and auditing of AI systems used in the customs clearance procedure, in order to provide, at the same time, customs efficiency and respect for the fundamental rights and guarantees of those administered, users of customs services.