FACIAL RECOGNITION TECHNOLOGY THROUGH A CRITICAL LENS: A REVIEW AND ETHICAL RISK TAXONOMY

Abstract

Facial Recognition Technology (FRT) has transitioned from a specialized research area to a pervasive societal force, yet its rapid advancement often obscures profound ethical and technical challenges. This review provides a critical synthesis of FRT, moving beyond descriptive accounts to analyse its trajectory, core methodologies, and societal entanglements. The authors trace the evolution from early geometric approaches to the current deep learning paradigm, critically examining the inherent trade-offs particularly the tension between accuracy gains and the entrenchment of bias, loss of interpretability, and escalating privacy risks associated with methods like CNNs, embedding networks, and emerging 3D or multimodal systems. Furthermore, the review examines the expanding use of facial recognition technology across sectors, weighing its claimed benefits against societal harms such as algorithmic bias, surveillance expansion, and data security risks. It shows that despite technical progress, FRT systems often fail in real-world settings and reinforce existing inequalities. A key contribution is the introduction of a Taxonomy of Ethical Risk to classify applications based on their potential for harm. The analysis emphasizes the urgent need for robust ethical frameworks and transparent governance. It calls for a shift toward privacy preserving and fairness aware design approaches. Overall, the review offers critical guidance for researchers, policymakers, and stakeholders on the responsible use of FRT