In an historical moment in which the myriad dimensions and consequences of a housing crisis characterising much of the world are at the forefront of policy and research, AI is emerging as a transformative force likely to rework multiple aspects of work, social life, cities, and, by implication, the housing system. Numerous institutions such as the World Bank, state and local governments, and housing providers are directly posing AI as a means to resolve the global housing crisis. The World Bank claims that “AI can address the global housing deficit” (Walley et al 2023) while, in Australia, former NSW Housing Minister Paul Scully argued that there is “zero per cent chance of meeting supply unless the government and sector immediately adopt AI technology” (Fuller 2023). Behind these claims is a belief that AI can target and quicken the processes of individual components of housing provision, from construction management to the processing of development approvals, to increase the amount of housing available.
In this presentation I illustrate a framework and approach through which to explore the impacts of the interweaving of housing and AI and enable researchers to apprehend the complexity of AI and housing. To do this we draw on, and move beyond, the rich housing literature on algorithms (and more recently AI) across the rental sector, property technology (proptech) and mortgage markets (Fields 2024; Nethercote 2023; Rogers et al 2024, Maalsen et al 2026; Ciočanel et al 2024; Zou and Khern-am-nuai 2022) in three ways.
First, we move beyond the focus on AI in individual sectors of the housing system to a focus across the multiple elements of the housing system, including but not confined to how housing is constructed, how it is financed, how it is allocated (eg tenure), how it is regulated, and how it is inhabited. Analysis across the housing system-as-a-whole will illuminate the variety of AI use in housing and importantly shed light on its cumulative and interrelated impacts.
Second, and following from the first, we move away from generalised “algorithm talk” (or in this case AI talk), given its lack of specificity to diverse algorithmic operations (Iapaolo and Lynch 2025; Amoore 2020, 244). Algorithms, including those that underpin AI are varied in their logics and grammars, architectures and domains, and they use domain-specific data (Cadman et al 2025). Because of this, generalising AI’s use and outcomes across housing is not particularly helpful. It is critical therefore that we conduct grounded case study research to understand the myriad of ways that AI and its associated algorithms are operating in and impacting on housing and our responses to this.
Third, we propose a methodological orientation that will enable us to open up new ways of thinking about the cumulative impacts as AI becomes embedded across housing systems and increasingly agentic in its operations. Advancing beyond conventional approaches we draw inspiration from post humanist and more-than-human approaches (e.g. Hayles 2017; Rose 2019; Maalsen 2023; Tiwari 2025; and Iapaolo and Lynch 2025) to position AI as collaborators in producing the housing system. This conceptual reframing also opens up novel ways for intervention and resistance to the possible negative impacts of AI in housing.
We illustrate this through interviews with planners involved in an Early Adopters pilot of AI in the state of New South Wales, Australia, where we explore the work required to make AI suitable for planning use. Algorithmic collaboration is far from a seamless partnership. It only becomes possible through the labour involved in ensuring data readiness, testing and iterating, advocacy and translation work required to ready institutions for AI, and, ideally, collaboration across geographical and institutional contexts. This labour inevitably complicates both the tech industry and government policy rhetoric that AI will speed up planning processes, improve productivity and reduce the administrative burden on planners and deliver more housing (NSW Gov n.d). Our work identifies the interweaving of planning and technological skills and draws out the key dimensions of organisational, educational and culture change required to proactively shape beneficial AI integration. Such processes have implications for the uptake of AI within planning and this has implication for our cities, more broadly
Presenter: Sophia Maalsen
Research team: Sophia Maalsen, Pauline McGuirk, Robyn Dowling, Claire Daniel
