Control and Resistance in Automated Shops: Retail Transparency, Deep Learning, and Digital Refusal

Abstract

Through the enrolment of big data, deep learning, sensor fusion, and computer vision technologies, Amazon Go and similar shops pursue the automated management of retail subjects, goods, and transactions. Tracing the logics of automated shop technology, the paper makes two contributions. First, it proposes a theory of “retail transparency” to attend to how automated shops reimagine space as a series of pockets of excess (actions that escape circuits of capitalist valuation) to be countered through acts of making-transparent (datafication for integration into digital systems of control). Retail transparency is underpinned by interventions aimed at perceiving, incorporating, and productivising excess. Second, we argue that logics of deep learning raise important challenges to traditional conceptions of resistance in digital geographies, as these tend to rely on a celebration or cultivation of excess. Instead, we offer a speculative reflection outlining a politics of “circuit-breaking” which refuses to engage algorithmic logics on their own terms.