【摘要】 在“双碳”目标与制造业智能化转型背景下，选取2010—2023年我国制造业上市公司数据，通过文本分析法量化人工智能应用水平，实证检验人工智能应用对能源强度的作用机制及影响效应。研究发现，人工智能应用显著降低企业能源强度，结论经内生性处理和系列稳健性检验后仍然稳健。机制分析表明，人工智能应用通过提升生产效率、降低管理成本及优化能源使用效率三重路径抑制能源强度。异质性分析显示，该效应在非国有企业、大规模企业及非智慧城市试点企业中更为显著。研究突破传统测度局限，构建多维人工智能指标，揭示人工智能技术抑制能强度的“黑箱”，为制造业智能化转型与能源政策优化提供理论支撑和实践依据，助力经济高质量发展。更多还原

【Abstract】 Against the backdrop of the "dual carbon" goals and the intelligent transformation of the manufacturing industry,drawing on data from listed Chinese manufacturing firms spanning 2010 to 2023,this study quantifies the level of artificial-intelligence adoption through textual analysis and empirically examines how AI affects energy intensity and the channels underlying this relationship. The results show that AI adoption significantly lowers firm-level energy intensity,and this finding survives endogeneity corrections and a battery of robustness checks. Mechanism tests reveal that AI curbs energy intensity via three routes:Raising productive efficiency,cutting managerial costs,and improving energy-use efficiency. Heterogeneity analyses indicate that the effect is stronger for non-state-owned,large-scale,and non-smartcity firms. By moving beyond conventional measurement approaches and constructing a multidimensional AI index,the study opens the "black box" of how AI technology restrains energy intensity. The findings offer both theoretical insights and practical guidance for the smart transformation of manufacturing and the refinement of energy policies,supporting high-quality economic development under China’s "dualcarbon" goals.更多还原