AI Assists Urban Planning, But Safeguards Needed

Artificial intelligence is increasingly being used in urban planning to improve flood mitigation, generate neighborhood renderings, and enhance public communication, according to a recent study published in Nature by Northeastern University professors Esteban Moro and Ryan Wang. The research analyzed over 100 studies in urban science, computational social science, and geospatial AI, focusing on two types of AI: distribution-fitting generative models and foundation models.

Distribution-fitting generative models can create new images based on large datasets, such as generating city layouts from satellite data. These models are used to predict flood scenarios by learning from global flood data. Foundation models, which rely on human language, are being used to simplify complex policy documents and even represent community perspectives in planning meetings. One study showed AI agents influencing land development plans in Beijing by voicing resident concerns.

Despite the benefits, the researchers warned of risks, including biased outputs and the potential for misinformation. They emphasized the need for governance frameworks to ensure fairness, privacy, explainability, and community involvement. Urban planners must balance AI's efficiency with human oversight to avoid ethical and technical pitfalls.