Artificial Intelligence (AI) has begun to permeate urban planning in increasingly diverse ways. Its applications are no longer limited to the efficient management of resource-based urban systems (such as traffic, utilities, security); increasingly it is used to generate urban design and planning options across diverse domains, including spatial forms, scenario sets, and preliminary planning assessments. Embedded in planning processes, AI tools can therefore shift the activity itself: from reacting to problems as they surface towards anticipating them, in a more proactive mode of city governance.

In this context, tools that automatically and procedurally generate urban forms based on human input (typically “prompts”) are becoming commercially mature and will undoubtedly be adopted by the public and private sectors. A systematic review of 120 generative urban design studies confirms that generative methods, from evolutionary optimization to deep generative models, are already used to explore complex solution spaces and produce design options satisfying conflicting objectives and constraints at the early design stage (Jiang et al., 2024).

AI is thus moving from the operational back office into the normative and imaginative spaces where urban decision-makers and municipalities decide what should be built, preserved, densified, or transformed across existing silos and domains.

For the UNECE Committee on Urban Development, Housing and Land Management therefore, the core question shifts from whether AI will affect planning, to the institutional, ethical, economic, aesthetic and technical conditions under which it can improve urban outcomes, while reinforcing participation, engagement and equity in urban strategy, planning and design.

This paper presents the collective reflections of three Genva UN Charter Centres of Excellence (CoEs) on Sustainable Housing and Urban Development on this topic, emphasizing the challenges and opportunities ahead for AI use in Urban Planning. Our paper makes no claim for completeness nor research precision rather, based on our individual and collective practice, research and granular experience it aims to help inform the Committee on potential areas of intervention before they become critical challenges