‘Assisted’ facial recognition and the reinvention of suspicion and discretion in digital policing[‘Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification’]

Automated facial recognition (AFR) has emerged as one of the most controversial policing innovations of recent years. Drawing on empirical data collected during the United Kingdom’s two major police trials of AFR deployments—and building on insights from the sociology of policing, surveillance studies and science and technology studies—this article advances several arguments. Tracing a lineage from early sociologies of policing that accented the importance of police discretion and suspicion formation, the analysis illuminates how technological capability is conditioned by police discretion, but police discretion itself is also contingent on affordances brought by the operational and technical environment. These, in turn, frame and ‘legitimate’ subjects of a reinvented and digitally mediated ‘bureaucratic suspicion’.