Development or dystopia? : an introduction to the accountability challenges of data processing by facial recognition technology

Abstract

In parallel with the increasing deployment of Facial Recognition Technology (FRT), scholars are starting to pay attention to the inherent legal issues. We can identify a wide body of literature pointing out fairness concerns regarding the training and performance of these systems. Moreover, there is a scholarship stream that focuses on the threat that FRT might pose for the right to privacy in general and to data protection, in particular. However, most of these works do not provide an overview of the technology to prevent misunderstandings about how FRT works and what it can and cannot achieve. This analysis is necessary because it will inform a realistic and constructive debate around the legal implications of FRT deployment. For instance, the lack of consensus among sources when defining this technology, both from hard and soft law instruments, might reflect that different and separate stages of the actual process, or different functions, are being taken as FRT. Furthermore, although accountability has been spotted as a key feature to address the abovementioned fairness or privacy issues, there are no studies on how, and on what points through the FRT pipeline, accountability might arise. This results in positions ranging from risk assessment to extreme solutions such as banning.