AI in the street: scoping everyday observatories for public engagement with connected and automated urban environments

In recent years, the street has emerged as one of the primary sites where everyday publics encounter AI. Industry and public sector organisations have undertaken street trials of AI-based technologies ranging from automated vehicles (AVs) and facial recognition (FR) to data-driven modelling supporting smart city monitoring and mobility systems. Such street-level AI deployments have given rise to policy initiatives across local, regional and national levels defining societal benefits of AI innovation (safety, sustainability, inclusion, levelling up, etc.) as well as efforts at institutional engagement with affected local communities through policy exhibitions, user-centred design workshops (Philips, 2020) and citizen cafe’s (Jennings et al, 2020). However, at the same time, street trials of AI-based technologies have provoked citizen contestations that significantly exceed existing engagement frameworks, highlighting public concern with surveillance, discrimination, access, and civil liberties. AI trials in the street have also surfaced divides between the general discourse of responsible innovation that frames national investments in AI and the everyday realities and concerns of life in the street. In order to build capacity across local communities and AI innovation partners to bridge this divide, this project will evaluate, trial, prototype and translate a specific situated method for public engagement with AI: street-level observatories of everyday AI.

This project’s overarching aim is to build on the classic Mass Observation Methodology to situate public engagement with AI at a human scale by grounding it at a primary site where the transformations, potential benefits, harms and (ir)responsibilities of AI in society can be made visible and explorable for everyday publics. To this end, we will 1) build partnerships across the arts, social research and computer science and with organisations and groups invested in situated forms of public engagement with innovation in the broad area of connected and automated mobility. 2) trial street-level observatories of everyday AI at 5 sites across the UK and Australia. Together with non-academic partners we have devised a design-led process to evaluate, trial and develop existing methods in this area, with try-outs of everyday AI observatory methods in different cities at the heart of the project: Cambridge, Coventry, London and Edinburgh. City locations, observatory methods and thematic foci of the observatories have been selected to ensure our scoping exercise has sufficient breath while ensuring comparability, and will combine digital and place-based approaches. To further expand our space of comparison, we have included one international test case, an observatory of drone-based trials in “streets in the sky” over the largely sub-urban city Logan (Australia).

Our trials of situated AI observatory trials will set the stage for 3) a joint process of prototyping and translation of an everyday AI observatory, with a special focus on the design of everyday AI observatories that is effective as a relational methodology, realising the double purpose of making hidden AI-based infrastructures visible for everyday publics, and making everyday public understandings of AI in society visible to expert communities. To support the prototyping process, we will conduct three translation workshops with relevant communities in science and industry, public policy and with citizen groups. Our overarching aim, in doing so, is to trial and develop a methodology that can help the wider bridging of, on the one hand, the stakes and issues that preoccupy the tech sector, industry, and, increasingly, the national government, in the framing of responsible AI on the national level, and, on the other hand, what matters to those who live on the streets of the nation.