FACIAL RECOGNITION, DUE PROCESS, AND ALGO, AND ALGORITHMIC RACISM: TECHNOLOGICAL FLAWS AND WRONGFUL CONVICTIONS IN BRAZIL

Abstract

This article aims to critically analyze the risks associated with the use of facial recognition in the Brazilian criminal justice system—particularly regarding the occurrence of wrongful convictions—in light of due process and constitutional criminal law principles. Starting from the premise that the technology, while useful for investigations, is fallible and can exacerbate existing inequalities, the study discusses how human, technological, and institutional failures—compounded by the phenomenon of algorithmic racism—contribute to the arrest and conviction of innocent individuals. A qualitative methodology is employed, involving bibliographic and documentary research based on legal doctrine, legislation, recent court rulings (including STJ Repetitive Theme No. 1,258), and reports from institutions such as Public Defender's Offices and wrongful conviction review projects. Ultimately, the article seeks to identify legal and institutional parameters for the responsible use of facial recognition—aligned with the Federal Constitution and criminal procedural guarantees—in order to minimize injustices and violations of fundamental rights.