Integrating artificial intelligence for sustainable urban planning and architectural development: A framework for smart and resilient cities

The rapid urbanization worldwide and intensifying climate change impacts call for urgent, transformative approaches to city planning and the design of built environments. Conventional urban planning techniques, which rely on deterministic models and expert heuristics, are increasingly insufficient for addressing the complex, multifaceted sustainability issues confronting modern cities. This paper introduces the AI-SUPAD framework a six-layer hierarchical decision-support system that integrates machine learning, deep learning, reinforcement learning, and generative AI into a comprehensive analytical platform. This platform facilitates urban-scale planning and architectural development. Using a mixed-methods research approach, the framework was trained and tested on a realistic simulated dataset comprising 1,008 zonemonth observations across 12 urban zones and three synthetic city types (2019–2025). The dataset included variables such as population density, energy demand, traffic congestion, green space, and per-capita carbon emissions. Five AI algorithms Random Forest, XGBoost, CNN, LSTM, and the AI-SUPAD ensemble were benchmarked using standard performance metrics. AISUPAD outperformed the others, achieving a predictive accuracy of 94.3%, an F1-score of 0.944, and an R² of 0.962. Global smart city case studies Singapore, Barcelona, Amsterdam, Masdar City, and NEOM showed postdeployment reductions: 47.3% in carbon emissions, 32.3% in energy use, and 79.7% in green space. The framework supports UN Sustainable Development Goals 7, 9, 11, 12, 13, and 15. These results provide evidencebased insights for urban planners, architects, policymakers, and smart city professionals worldwide