Artificial Intelligence, Urban Design and Integrative Planning: The Potential Benefits and Risks of AI-Driven Urban Planning

Artificial intelligence (AI) is moving urban planning from reactive administration toward anticipatory, generative and integrative forms of governance. Yet its growing use in scenario testing, spatial analysis and design generation raises questions concerning institutional capacity, equity, cultural pluralism and democratic accountability. This paper examines the conditions under which AI can support more sustainable, resilient and participatory urban planning without displacing human judgement or weakening municipal sovereignty. Drawing on the combined practice and research of three Geneva UN Charter Centres of Excellence, it positions AI within systems-oriented urban governance rather than treating it as an autonomous planning technology. The discussion explores how explainable AI, geospatial data, digital twins and continuous feedback can reveal interdependencies among housing, transport, health, education, landscape, energy and climate systems. Such tools may lower analytical costs, broaden access to technical capability and help smaller municipalities compare alternative development trajectories. However, the paper also identifies three major risks. First, routine reliance on AI may erode planners’ synthetic, interpretive and imaginative capacities, reducing professionals to evaluators of machine-generated options. Second, models trained on unevenly distributed and culturally specific data may privilege dominant urban forms while marginalizing indigenous, informal and non-Western spatial practices. Third, proprietary platforms, recurring computational costs and restricted auditability may produce vendor lock-in and new dependencies, particularly for resource-constrained municipalities. The paper therefore argues that AI should augment rather than replace public planning capacity. A responsible framework requires five governance principles: support rather than substitution, institutional capacity before scale, plural datasets and contextual evaluation, open standards with credible exit options, and meaningful participation and accountability. Procurement, regulation and professional education must consequently be treated as core governance functions. The decisive challenge is not whether AI can plan cities better than humans, but whether municipalities can mobilize it while preserving critical judgement, cultural diversity and durable institutional autonomy.