Machine Vision and Tagging Aesthetics: Assembling Socio-Technical Subjects through New Media Art

Abstract

This paper builds on the concept of ‘tagging aesthetics’ (Bozzi, 2020b) to discuss new media art projects that combine machine vision and social media to address how different kinds of socio-technical subjects are assembled through AI. The premise outlines how the naturalisation of machine vision involves a range of subjects, juxtaposed along different conflictual lines: ontological (human-machine), biopolitical (classifier-classified), socio-technical (tech worker-data cleaner), political (AI-viewing public). Embracing the ambiguity inherent in the shifting boundaries of these subjects, I analyse works by different new media artists who approach one or more of these juxtapositions by engaging with diverse forms of tagging. The practice of tagging is often discussed through data-driven analyses of hashtags and how related publics can be mapped, but in my framework, tagging can encompass a wider spectrum of techno-social practices of connection (e.g. geotagging, tagging users). I discuss artworks by Kate Crawford and Trevor Paglen, Dries Depoorter and Max Dovey to illustrate how these practices can be leveraged artistically to make visible and even ‘stitch together’ the manifold subjects of machine vision. I explain how those taggings denaturalise processes of socio-technical classification by activating awareness, if not agency, through the sheer proximity they enact. Far from being a tool to map knowledge and essentialised identities, tagging aesthetics are ways to perform the techno-social and shape future cultural encounters with various forms of others. By exploring different approaches to tagging aesthetics – (dis)identification, semi-automated assembly and embodied encounter – this paper illustrates how tagging can be used to culturally negotiate the impact of machine vision in terms of issues such as surveillance and the performance of digital identity.