Artificial intelligence (AI) is a general-purpose technology driving technological and industrial transformation. In this study, we construct an AI adoption indicator based on textual analyses of firms’ annual reports and measure innovation quality using the cumulative three-year forward citation counts—excluding self-citations—of invention patents at the firm-year level. Based on a sample of Chinese A-share manufacturing firms spanning 2011–2021, we found that AI adoption is positively associated with innovation quality. This conclusion remains robust across a range of endogeneity-mitigating specifications and robustness checks. Exploratory mechanism analyses suggest that this positive association is partly accounted for by three channels—R&D investment, human capital, and R&D efficiency—although these results should be interpreted as indicating an association rather than causal mechanisms. Heterogeneity analysis further showed that such association is concentrated in larger enterprises, and it is stronger for enterprises with lower financialization and those in asset-intensive industries. However, no significant difference was observed across firms with different ownership or regional locations. This study expands the literature on the innovation effects of AI and offers new insights into the factors influencing innovation quality in the AI era. It also has implications for business management and public policy with regard to leveraging AI to foster high-quality innovation capabilities, thereby providing a knowledge foundation for industrial upgrading and sustainable development.

In this study, we utilized multi-regional input–output (MRIO) tables from 2012 to 2017 to determine the spatiotemporal characteristics of CO₂ emissions, air pollutant emissions, and value added associated with trade in different regions and industries, as well as the level of coupling coordination among them. Secondly, structural decomposition analysis (SDA) was used to identify the drivers of changes in the above indicators at the regional level. The results show that consumption-based emissions exceeded those based on production in developed regions such as Jing-Jin and the eastern coastal regions, while the opposite occurred in energy hubs such as the northern coastal and central regions; the results of the value added show higher levels in production areas than in consumption areas in the eastern coastal regions, while the opposite trend was observed in the northwestern regions. In different industries, energy production and heavy manufacturing contributed significantly to CO₂, PM_2.5, and SO₂ emissions, while the service industry contributed more to NO_x and VOC emissions and value added. The relationships among the changes in the three consumption-based indicators were divided into four categories: positive synergies, negative synergies, trade-offs A, and trade-offs B. Recommendations for targeted collaborative management strategies were delineated based on a regional resource-driven classification.