本文基于2011-2019年中国30个省份的面板数据测度工业智能化指标，实证检验了工业智能化与碳生产率之间的影响机制。结果显示：工业智能化显著提高了碳生产率，这一结论在进行一系列稳健性检验后仍然成立。机制分析表明，工业智能化主要是通过多元化集聚间接提高碳生产率，由于产业集聚生命周期的存在，工业智能化对碳生产率的影响呈现出“先增后减”的非线性特征。空间计量分析表明，工业智能化不仅可以提高本地的碳生产率，还可以提高邻近地区的碳生产率，有助于形成地区间绿色协调发展的空间格局。异质性分析表明，在非资源依赖、工业化水平高、环境规制强度低的地区，工业智能化提高碳生产率的作用更显著。研究结果为提高碳生产率进而实现“双碳”目标提供了一条可行的路径，也为政府部门积极谋划区域工业智能化战略布局提供了有益的借鉴。

Industrial intelligence with "intelligent manufacturing” as the core has provided a crucial driving force for China to achieve the goal of "carbon peak and carbon neutrality." As the world's largest emerging economy, China is accelerating the construction of emerging infrastructure such as 5G, cloud computing, and artificial intelligence. At the same time, as global warming continues to accelerate, China stresses the need to protect the global environment and achieve a harmonious coexistence between man and nature. In 2021, for the first time, China included "carbon peak and carbon neutrality" in its government work report. The vigorous development of industrial intelligence characterized by "machine for man" can not only reduce production costs of enterprises to improve production efficiency through the application of intelligent machines but also reshape economic geographic patterns by changing factor endowment conditions, realize resource sharing among enterprises, and reduce carbon emissions. Therefore, studying the relationship between industrial intelligence and carbon productivity and its mechanism is of great theoretical and practical significance. Based on the panel data of 30 provinces in China from 2011 to 2019, this paper measures the industrial intelligence indicators from three aspects: the basic conditions of intelligence, the degree of intelligent application, and the achievements of intelligent technology, and empirically tests the spatial-temporal evolution characteristics and influence relationship between industrial intelligence and carbon productivity. This paper includes industrial agglomeration into the research framework and expounds on the nonlinear relationship between industrial intelligence and carbon productivity based on the life cycle theory of industrial agglomeration to analyze the influencing mechanism between industrial intelligence, industrial agglomeration, and carbon productivity. In addition, this paper also takes resource dependence, industrialization level, and environmental regulation intensity into consideration to explore whether it has a heterogeneous impact on the relationship between industrial intelligence and carbon productivity, and improve the theoretical framework of the relationship between industrial intelligence and carbon productivity. The results show that industrial intelligence significantly improves carbon productivity. And this conclusion remains valid after a series of robustness tests including the introduction of instrumental variables, the replacement of the regression method, and the replacement of the calculation method of explanatory variables. Mechanism analysis shows that industrial intelligence promotes diversified agglomeration and specialized agglomeration. But it mainly indirectly improves carbon productivity through diversified agglomeration. Due to the existence of the life cycle of industrial agglomeration, the influence of industrial intelligence on carbon productivity presents a nonlinear feature of "first increase and then decrease." Spatial econometric analysis shows that industrial intelligence has a spatial spillover effect. Industrial intelligence can improve carbon productivity not only in local areas but also in neighboring areas, which will be conductive to shape a spatial pattern of coordinated green development among regions. Heterogeneity analysis shows that the role of industrial intelligence in promoting carbon productivity is more significant in regions with non-resource dependence, high industrialization levels, and low environmental regulation intensity. The results of this study provide a feasible path to increase carbon productivity and achieve the "double-carbon" goal and also provides a beneficial reference for government departments to plan the strategic layout of regional industrial intelligence energetically.

Artificial intelligence (AI) is the core technology of digital economy, which leads the transition to a sustainable economic growth approach under the Chinese-style environmentally decentralized system. In this paper, we first measured the green total factor productivity (GTFP) of 30 Chinese provinces from 2011 to 2020 using the super-efficiency slacks-based measure (SBM) model, analyzed the mechanism of the effect of AI on GTFP under the environmental decentralization regime, and secondly, empirically investigated the spatial evolution characteristics and the constraining effect of the impact of AI on GTFP using the spatial Durbin model (SDM) and the threshold regression model. The findings reveal: a U shape of the correlation of AI with GTFP; environmental decentralization acts as a positive moderator linking AI and GTFP; the Moran index demonstrates the spatial correlation of GTFP; under the constraint of technological innovation and regional absorptive capacity as threshold variables, the effect of AI over GTFP is U-shaped. This paper provides a useful reference for China to accelerate the formation of a digital-driven green economy development model.