Reciprocity Deficits: Observing AI in the street with everyday publics

The street has emerged as a primary site where publics are con-fronted with AI as an everyday reality. AI-based systems are nowcommonly deployed in this setting in the form of automated cars,facial recognition, smart billboards and the like. While this use of AIin the street has attracted significant media attention and publiccontroversy, such deployments often remain inscrutable, andeveryday publics are often unaware of their presence. We reporton participatory engagements with AI in the situated environmentof city streets under these paradoxical conditions. Combiningapproaches from social and cultural studies of AI, design researchand Science and Technology Studies (STS), we engage with thestreet as a site for ‘material participation’ in AI through what we calleveryday AI observatories. We narrate and reflect on participatoryinterventions at these observatories, conducted in five cities in theUK and Australia, and highlight three ways of understanding AI thatemerged from them: (1) the street as a transactional environment,(2) the designed invisibility of AI in the street, and (3) a perceivedstratification of street environments through statistical governance.We conclude by reflecting on the reciprocity deficits observedbetween AI infrastructures and everyday publics in the street.