Radical Spatial Restructuring: Planning With AI, Public Discontent, and the Politics of Emerging Spatial Imaginaries

Planning for future spatial and land-use change represents a central challenge of the green transition (Concito, 2023), yet the societal and environmental implications of such planning remain only partially understood (Blok, 2025). In Denmark, the recent green tripartite agreement exemplifies this tension. While the agreement seeks to accelerate renewable energy, infrastructure development, and spatial restructuring, it has also contributed to increasing “green discontent.” Local communities have expressed fatigue, mistrust, and frustration regarding decisions perceived as opaque, topdown, or socially uneven (Stahl, Ellersgård, & Mulvad, 2025; Bourdin, Molica & Marques Santos, 2025). In this context, generative artificial intelligence is increasingly promoted as a tool to visualise and communicate future spatial transformation of neighbourhoods, landscapes, and of everyday life (Duray, Gulsrud, Burgos-Thorsen, forthcoming; Hermund, Ribeiro, & Gjorgjievski, 2025). Images of future scenarios have the capacity to shape our deeply held, collective understanding of socio-spatial futures. The concept of “spatial imaginaries”, defined as socially embedded manifestations (Davoudi et al., 2018; Davoudi & Machen, 2022), is used in this paper to discuss the potential of generative AI in articulating common aspirations and identities. By engaging with the concept of spatial imaginaries, the authors critically discuss how these technologies operate within broader infrastructures of techno-capitalism that influence not only what can be planned and visualized, but
also which futures are perceived as legitimate, desirable, or even conceivable (Raymond et al., 2025).

This paper conceptualises generative artificial intelligence as both an infrastructure of imagination and an agent of infrastructural imagining (Gandy, 2022; Graham & Marvin, 2022; Cugurollo et al., 2024). The analysis critically examines how globally trained models, shaped by partial datasets and platform logics, tend to reproduce dominant “sustainable city” aesthetics. This process often sidelines situated identities, ecologies, and social values that are essential for navigating Denmark’s
contested green transition (Raymond et al., 2025; Klein & D’Gnazio, 2024; Gulsrud, 2018; Gulsrud et al., 2018). In discussing the use of generative artificial intelligence in planning processes, this paper discusses how representational biases risk amplifying existing discontent by providing abstractions rather than recognising lived experiences, territorial attachments, or local forms of knowledge
(Williams et al., 2024; Gulsrud et al., 2018).

To address these challenges, this paper contends that planning with artificial intelligence must explicitly engage diverse publics through participatory, community-rooted practices that prioritise transparency, plurality, and democratic authorship. Artificial intelligence should not serve to accelerate predetermined visions, but rather as a democratic tool that facilitates new spaces for contestation, negotiation, and collective imagination. This approach is currently being explored through a
prototype interactive installation in the exhibition “Imagining the Future” at the Royal Danish Academy, where audiences interact with local images and co-create future scenarios using generative AI (Hermund, Ribeiro, & Gjorgjievski, 2025). In Denmark and internationally, fostering reflexive planning processes that include community-rooted spatial imaginaries is essential for rebuilding trust and enabling more just and inclusive green transitions.