Situated within efforts in urban studies to move beyond smart city paradigms, Artificial Intelligence and the City offers a timely interdisciplinary contribution to comprehending how AI is (re)shaping contemporary urban conditions. As AI proliferates across, within and beyond the city, becoming embedded in and productive of the assemblages which shape urban life, this book resists treating it as monolithic, instead highlighting the heterogeneous forms, capacities and presences through which AI becomes situated and consequential in urban space. Contributors explore a spectrum of manifestations, from robots and autonomous vehicles to city brains and urban software agents, spanning highly visible systems to embedded algorithmic processes that subtly shape urban rhythms. Collectively, these systems are framed as contributing to ‘AI urbanism’, a distinct condition emerging from the diffusion of ‘artificial intelligences’ throughout the contemporary city.

In unpacking AI’s distinctive nature, the book argues that smart city frameworks are increasingly insufficient for grasping AI-enabled transformations. The editors propose the framework of ‘function’, ‘presence’ and ‘agency’ to capture these shifts, illustrated across the volume’s case studies, tracing a movement from reactive systems of measurement toward more anticipatory modes of operation, alongside AI’s growing visibility within shared urban environments. Smart et al.’s analysis of ‘City Brain’ systems in China, for example, illustrates how AI is increasingly embedded within expanding digital urban infrastructures, enabling real-time analysis and the proactive deployment of resources through access to extensive governmental databases and the continuous collection of urban data. Meanwhile, Jackman’s examination of drones highlights the growing presence of aerial technologies in everyday urban environments. Agency is positioned as a reconfiguration of decision-making logics under conditions of uncertainty, equally reflected in chapters on autonomous vehicles that must continually adapt to the unpredictable contingencies of urban contexts.

Taken together, the chapters advance another central claim—that AI and the urban condition are fundamentally co-constitutive, showing how AI depends on urban environments, from material infrastructures and existing data landscapes to situated processes of learning and adaptation as systems operate ‘in the wild’.  For instance, Lynch and Del Casino’s chapter argues that the ‘intelligence’ of urban AI is fundamentally relational, emerging through encounters with particular urban places and practices. In turn, AI co-produces new epistemologies of the city, transforming how urban space is made legible and ultimately acted upon. As Sumartojo (p. 163) emphasizes, this situates AI as always in relation to the people and environments in which it operates, complicating claims that it functions ‘autonomously’ in urban space.

Central to this relational framing is the role of data. AI urbanism is sustained by dense data ecologies that enable systems to interpret, predict and intervene in urban processes, with urban life generating the continuous data streams that sustain AI. For instance, Sweeney’s analysis of domestic technologies illustrates how everyday interactions with devices become folded into expansive data assemblages extending far beyond the home, whilst also consolidating the power of platform actors and technology companies. Other chapters trace similar dynamics, from data-driven property valuation algorithms shaping housing markets and facilitating real-estate investment (Chapter 17) to the growing influence of platform companies through data ecosystems that continuously learn from and coordinate urban behaviour at scale, as seen with services such as Uber (Chapter 12). Across these cases, the data underpinning urban AI emerges as deeply embedded within broader socio-technical networks of capital and power.

Yet these data ecologies are inseparable from the infrastructures that sustain them. The book effectively elucidates how AI manifests materially in the city, and how data translates the socio-material dynamics of urban life into AI systems, but the material impacts of the computing infrastructures that sustain urban AI remain comparatively underdeveloped. AI relies on extensive physical infrastructures, including high-performance computing architectures and hyperscale data centres, which carry significant spatial and ecological implications, from the contested siting on urban peripheries to the energy demands associated with training and operating large-scale machine learning systems. As debates surrounding the environmental costs of AI infrastructures intensify, greater engagement with these material geographies would further strengthen the book’s socio-technical approach by highlighting the environmental and territorial footprints of urban AI systems.

Writing during my PhD fieldwork on AI in the home, I found the book’s recurring attention to the everyday particularly resonant. While the chapters trace AI’s emergence across multiple scales, Sumartojo’s call to ‘attend to how AI urbanism articulates in normal life, and at smaller scales than the whole city’ (p. 160) feels especially generative, prompting questions about how to account for the largely invisible but profound consequences (p. 15) of these systems. Contributions by Jackman and Sweeney are central here, showing how such dynamics take shape within urban domestic settings and how local encounters connect to wider relations of power. Similarly, early observations from my fieldwork point to the messy and uneven ways AI materializes across everyday domestic lives. From this perspective, the home emerges as a key everyday arena of urban AI, where greater attention could deepen understanding of the lived consequences of these technologies while further illuminating their entanglement within broader sociotechnical assemblages.

I also see value in further extending the book’s relational framing through more explicitly multi-actor approaches that continue to treat AI as an ‘overt sociotechnical phenomenon’, aiming to unravel the assemblages through which these systems emerge, evolve and materialize. Future work might move beyond singular urban sites or actors to ‘follow’ AI systems through everyday encounters, but also across their wider networks, from corporate constellations of actors (p. 45) who develop and govern to the often-overlooked gig labour that underpins their operation. Such approaches would deepen understandings of the co-constitutive dynamics identified throughout the volume, whilst supporting posthuman perspectives that foreground how machine intelligence emerges through situated interactions between humans, technologies and urban space.

Overall, the book offers a welcome intervention in debates surrounding urban AI and its uneven proliferation. It stands as a strong conceptual contribution, leaving me hopeful about future scholarship surrounding AI and its many incarnations, but also mindful of the need to remain critically engaged with these opaque systems and the complex sociotechnical transformations already reshaping urban life.

Alfie Greenwood, University College London

Federico Cugurullo, Federico Caprotti, Matthew Cook, Andrew Karvonen, Pauline McGuirk and Simon Marvin (eds.) 2023: Artificial Intelligence and the City: Urbanistic Perspectives on AI. London: Routledge.

Views expressed in this section are independent and do not represent the opinion of the editors.