Unmasking Trust: Examining Users' Perspectives of Facial Recognition Systems in Mozambique

Artificial Intelligence (AI) is being incorporated into a vast range of activities, changing how people relate to technology and with each other. However, most attention is focused on its technological aspects and disregards broader social consequences of the system implementation. This study investigates the factors influencing users' propensity to trust Facial Recognition Systems (FRS), a high-risk AI, in Mozambique. It employs a mixed methods approach, combining a survey involving 120 participants and 13 semi-structured interviews. The findings reveal that users' perceptions of FRS' robustness and principles of use significantly impact their trust levels. Moreover, this relationship is moderated by the effectiveness of system attribute communication and external issues. The outcomes of this study shed valuable light on crucial aspects that demand attention during the development of AI systems and foster reflection about existing issues, aiming to ensure the establishment of adequate levels of user trust.