CHALLENGES IN THE USE OF FACIAL RECOGNITION IN VIDEO SURVEILLANCE SYSTEMS IN BRAZIL

Abstract

The use of facial recognition systems in video surveillance has gained prominence in Brazil, being widely used in areas such as public security, transportation, and access control. However, its implementation faces significant challenges related to error rates and algorithm reliability, especially in complex social and infrastructural contexts. This article discusses the main limitations of these systems in Brazil, focusing on the factors that influence errors, such as image quality, algorithmic bias, and underrepresented demographic characteristics in databases. In addition, it addresses technical reliability, exploring performance metrics and technological constraints in singular scenarios. The text also highlights ethical and legal issues, such as the absence of specific regulations and the risks to privacy and the fundamental rights of citizens. Practical examples and Brazilian case studies are presented, highlighting the operational difficulties and social impacts of false positives and negatives. Finally, potential solutions are analyzed, such as advances in machine learning, training algorithms with diverse data, and the urgent need for robust legislation. The article concludes that, for facial recognition to become a reliable and ethical tool in Brazil, it is essential to invest in infrastructure, research, and regulation, ensuring a balance between technological innovation and the protection of fundamental rights.