Ding Wenhao, Hu Pan
School of Economics and Management, Northeast Agricultural University, Harbin City, Heilongjiang Province, China.
Sci Rep. 2025 Jul 2;15(1):22906. doi: 10.1038/s41598-025-06461-4.
The improvement of GTFP has become an important guarantee to change China's development mode and achieve long-term stable economic growth, while the development of AI is of great significance to the development of GTFP. This paper adopts the SBM-GML method to calculate the GTFP of 30 regions in China from 2011 to 2020, and manually collects AI data to study the impact of AI on GTFP, and systematically analyze the relationship between AI and GTFP using a panel double-fixed model with a faceted threshold model. The experimental results show that: (1) There is a positive and significant effect of AI on GTFP, which increases GTFP productivity by 0.3654% for every 1% increase in AI, which still holds after a robust type test. (2) The development of AI in China's eastern region has a greater promoting effect on GTFP. (3) Further mechanism analysis reveals that RIS and ER are two important channels through which AI influences the improvement of GTFP. (4) Threshold regression shows that AI has a single threshold effect on GTFP based on TI and LNPGDP. The promotion of GTFP by AI is higher when technological innovation is less than the threshold value of 29.95, and the promotion of GTFP by AI is insignificant when the level of economic development is less than the threshold value of 85.45. This paper deepens the knowledge and understanding of the role played in the development of AI at the macro level and provides suggestions for improving GTFP at the provincial level in China.
绿色全要素生产率（GTFP）的提升已成为转变中国发展模式、实现经济长期稳定增长的重要保障，而人工智能的发展对GTFP的发展具有重要意义。本文采用SBM - GML方法计算了2011 - 2020年中国30个地区的GTFP，并人工收集人工智能数据研究人工智能对GTFP的影响，运用面板双重固定模型和分面阈值模型系统分析人工智能与GTFP之间的关系。实验结果表明：（1）人工智能对GTFP有正向且显著的影响，人工智能每提高1%，GTFP生产率提高0.3654%，经过稳健性检验后依然成立。（2）中国东部地区人工智能的发展对GTFP的促进作用更大。（3）进一步的机制分析表明，研发投入存量（RIS）和经济回报（ER）是人工智能影响GTFP提升的两个重要渠道。（4）阈值回归表明，基于技术创新（TI）和人均地区生产总值（LNPGDP），人工智能对GTFP具有单一阈值效应。当技术创新小于29.95的阈值时，人工智能对GTFP的促进作用较高；当经济发展水平小于85.45的阈值时，人工智能对GTFP的促进作用不显著。本文深化了对人工智能在宏观层面发展中所起作用的认识和理解，并为中国省级层面提高GTFP提供了建议。
