Facial Recognition Technology in the Global Divide: Why Real-Time Biometric Identification Should Be Banned Beyond the EU AI Act

Abstract This article explores the global divide in regulating facial recognition technology for law enforcement, focusing on tensions between the European Union’s exceptions and the right to non-discrimination. Moving beyond privacy-based critiques, it grounds the analysis in equal treatment guarantees. Using Brazil as a case study, the article argues against reproducing the EU AI Act’s exceptions permitting real-time remote biometric identification by law enforcement. It demonstrates that even technically accurate FRT systems, when deployed within contexts of racialized policing and structural inequality, inevitably generate discriminatory outcomes. By highlighting the role of scale and limited contestability of algorithmic decisions, the article shows how real-time biometric identification produces a qualitatively distinct form of discriminatory harm. It concludes that Brazil, and countries facing similar structural disparities, should adopt an outright ban on real-time remote biometric identification for law enforcement, recognising that certain technological practices are inherently incompatible with constitutional guarantees of equality and non-discrimination.