Computing the face: From coloniality to control

Abstract Facial recognition will be integrated into EU biometric border control. In this article, we draw on theoretical insights from Deleuze and Guattari (2005), Browne (2010) and Amaro (2022) to construct a constellation of historical fragments of computing the face. We move away from a mere focus on bias in facial recognition algorithms and shift our attention towards the structural conditions that organise machine learning around the assumption of race. We unravel four fragments: departing from current proposals for the implementation of facial recognition in EU border management, we historicise computing the face through zooming into practices of physical anthropology and anthropometry in both the Belgian colonial context and by Bertillon and Galton where identification and classification intersected, and the technology of vision of early facial recognition systems. With our article, we show how computing the face has historically been intertwined with racialising instruments and effects and how current proposals on using facial recognition in EU border control are underpinned with a colonial epistemic gaze.