Reconhecimento facial na segurança pública dos estádios de futebol: estudo de caso da Arena Castelão e o desafio do viés algorítmico

Abstract

This study analyzes the challenges inherent in implementing facial recognition technologies in Brazilian public security, from the perspective of structural racism, personal data protection, and existing legal gaps. Beginning with a historical contextualization of public surveillance policies and the rise of the Surveillance State, the concept of technological racism is discussed, illustrated by algorithmic failures and problematic use cases of these tools in Brazil and abroad, which highlight risks and perpetuate inequalities. The issue of algorithmic bias in artificial intelligence and machine learning systems is investigated, contrasting it with the inadequacy of the current legal framework and draft legislation under discussion at various legislative levels. The study places particular emphasis on the obstacles to applying the General Personal Data Protection Law (LGPD) to facial biometric monitoring in football stadiums, addressing the processing of sensitive data, the complex definition of legal bases, the delicate balance between privacy and the public interest in security, the absence of specific regulation from the National Data Protection Authority (ANPD), and observed deficiencies in terms of governance, transparency, control, and accountability. To deepen the empirical analysis of the issues raised, a detailed case study is developed concerning the use of this technology at Arena Castelão, in Fortaleza/Ceará. Finally, it is argued that, despite the potential benefits for security, the indiscriminate adoption of such technologies entails significant risks of discrimination and violation of fundamental rights, demanding a qualified public debate, greater social control, and the development of specific and robust legislation to ensure its ethical, transparent, and responsible use.