Governing the housing system with AI: a framework

AI applications are rapidly emerging across the housing system but we know little about its use and impacts beyond generalised claims. Much of the discussion on AI and its impacts in housing remains generalised and unable to account for the diversity of AI applications, architectures and domains, as well as the sites in which it is being deployed. Without more specific knowledge it is difficult to account for the impact of AI in housing as well as identify effective ways to regulate its use. In this paper we develop and illustrate a framework and approach through which to explore the integration of AI and housing, to enable researchers to apprehend the complexity, trace the materialisation, and contemplate the impacts of the interweaving of AI technologies across the housing system. First, we argue for analysis across the housing system-as-a-whole to illuminate the variety of AI use in housing and importantly shed light on its cumulative and interrelated impacts. Second, we advocate a move away from generalised ‘AI talk’ to instead pay attention to the specificity and diversity of AI application. Third, we propose a methodological orientation that accounts for the dual agency of humans and AI technologies in housing outcomes. We argue that understanding AI’s impact calls for grounded empirical research that can account for its diversity and specifics across the housing system and suggest that housing researchers are well placed to do this work.