The impact of artificial intelligence on urban energy consumption

Managing urban energy use is a critical challenge for cities striving to achieve green and sustainable development, and also a vital focal point for propelling the synergistic growth between urban economic and environmental systems. Leveraging panel data from 281 prefecture-level cities in China spanning 2008 to 2021, this study adopts a two-way fixed effects (TWFE) model to assess the influence of artificial intelligence (AI) on urban energy consumption (UEC) optimisation. The results reveal that AI exerts a significant influence on the optimisation of UEC, and this impact stays robust following a series of robustness tests. Heterogeneity analysis, which accounts for geographical location and specific urban attributes, reveals substantial differences in the potential and actual effects of AI in optimising energy consumption across different cities. For cities in China's northeast and central regions, non-resource-based cities and low-carbon pilot cities, AI implementation has produced impressive results in the optimisation of energy consumption. However, for cities in eastern and western China, as well as resource-based cities and non-low-carbon pilot cities, AI's role in energy consumption optimisation has not been fully validated. The mediating effect test confirms that AI optimises UEC by enhancing green total factor energy efficiency. The findings have policy implications for the rational application and widespread deployment of AI in cities. The results serve as a basis for evidence-backed choices when developing precise energy policies, supporting the ongoing refinement of UEC and advancing the objective of sustainable urban development. • AI significantly optimises urban energy consumption (UEC) in Chinese cities through enhancing green total factor energy efficiency (GTFEE). • AI optimises UEC through three key mechanisms: dynamic supply-demand balancing, waste recycling, and real-time pollution monitoring. • Regional disparities exist: Northeastern/central cities show stronger AI-driven UEC optimisation than eastern/western regions. • Non-resource-based and low-carbon pilot cities achieve better energy efficiency improvements via AI applications. • Policy frameworks should integrate AI R&D, incentives, and region-specific strategies to address urban attributes.