Bridging or widening the divide? Exploring the role of AI technology in Smart City Development Efficiency and Spatial-temporal Divergence in China

Abstract To accurately assess the significance of AI in smart city development, this paper proposes an Entropy Meta Two-stage Dynamic SBM model and introduces the "EIT−6S" framework. The framework initially evaluates the development level of AI technology through Environmental Support, Industrial Development, and Technological Innovation (EIT), and assesses the efficiency of SC development based on six aspects: Smart Technology, Smart Economy, Smart Citizen, Smart Governance, Smart Environment, and Smart Transportation (6S). This study also introduces Kernel Density Estimation, the Gravity Model, and the Standard Deviation Ellipse Model to examine the temporal and spatial divergence characteristics of the AI system, SC system, and AISC system. The findings are as follows: (1) The average overall efficiency of smart city systems in 30 Chinese provinces was 0.86 from 2019 to 2023. It shows a slight downward trend. AI technology investment limits the efficiency of SC development and is the main source of inefficiency. The shortage of R&D funds and talent is the main bottleneck. (2) To improve the overall efficiency of AISC systems, policymakers should focus on improving the performance of the AI systems in Jilin, Zhejiang, and Hubei, preventing the continuous decline of efficiency in Tianjin and Liaoning, and focusing on the SC systems in Chongqing and Hubei. (3) The overall inefficiency in Northeast China stems from the widening differences in the AI technology base within the province, while the level of SC in the central and western regions drags down the efficiency improvement. (4) The development of AI technology is undergoing the process of "agglomeration followed by diffusion". The gravity centre of AI has a potentially dominant role in the formation and development of the gravity centre of AISC.