人工智能作为新一轮科技革命和产业变革的重要抓手，深刻影响着中国绿色发展的路径选择偏向。结合非连续性技术创新理论与中国现实情景，从产业结构优化视角构建人工智能技术赋能绿色增长的逻辑框架，并利用2010—2020年中国省级面板数据检验人工智能的绿色增长效应。结果表明，人工智能通过技术红利效应直接推动绿色经济增长，引入地区高校平均科技产出和《中国制造2025》政策冲击作为工具变量进行内生性修正后，人工智能的绿色增长效应仍显著存在。机制识别揭示，人工智能通过产业结构高级化和合理化驱动绿色经济增长，二者在人工智能绿色增长效应中的相对贡献分别为20.33%和8.35%。异质性分析发现，中国转型经济背景下，人工智能的结构红利在要素市场扭曲程度更低、创新人力资本水平更高、制度环境更完善的地区表现得更为明显，从而可以更充分释放其对绿色增长的赋能效果。拓展性分析发现，人工智能对绿色经济增长具有显著正向空间溢出效应，本地人工智能发展对空间关联地区的绿色发展绩效存在辐射带动作用。聚焦产业结构升级与绿色发展双重视角，可为塑造以人工智能为核心的技术竞争优势、实现经济高质量发展提供理论支撑和经验证据。

As innovation is the key of fostering green development which can meet people's aspirations for a "beautiful home" and a "beautiful China". Innovation in green technology is especially of great importance to achieve win-win results in economy, ecology and society in global green competition. At this stage, artificial intelligence (AI) which has the attributes of a new generation of general technology, strong spillover effect and high development potential, is undoubtedly the core force and important support to realize green technology innovation. The new generation of AI technology uses big data combined with strong algorithm power to gradually realize autonomous learning, rule identification and judgment decision, providing a new way to promote energy saving and consumption reduction. Therefore, the “14th Five-Year Plan” of the Chinese government focuses on strengthening fundamental research and original innovation,and improving the supply system of common basic technologies, and aims at the construction of a number of forward-looking and strategic national major science and technology projects such as AI.#br#The following questions remain unanswered in this context: does China's AI promote green economic growth? what is the transmission mechanism? what variances exist in the release of AI green impacts when diverse heterogeneous components are present? Furthermore, how can the government shape its AI development policies in order to fully capitalize on AI’s technological and structural benefits in green and low-carbon transformation?#br#To address these issues, this study builds an indicator system based on Chinese provincial panel data from 2010 to 2020 and implements modeling approaches such as the panel fixed-effects model, mediated-effects model, the moderated mediating effect model, spatial autoregressive model, and employs the instrumental variablesto empirically test the enabling effect, path mechanism, heterogeneous impact and spatial spillover effect of AI on China's green economic growth. Meanwhile according to the empirical test results, targeted industrial policies for the development of AI are put forward, so as to improve the quality and efficiency of the economy and achieve green growth.#br#The major conclusion is that AI drives green economic growth, and its green growth effect remains significant after accounting for endogeneity bias due to reverse causality and omitted variables. AI has a driving effect on green economic growth through an advanced and rationalized industrial structure,that is, AI releases technological dividends through direct effects and generates structural dividends by promoting industrial structure optimization. Moreover, in the context of China's economic transformation, the structural dividend of AI is more obvious in regions with lower distortion of factor markets, a higher level of innovative human capital, and a better institutional environment, and its empowering effect on green growth can thus be stronger. In addition, AI has a significant positive spatial spillover effect, and its development will drive green economic growth in economically connected regions.#br#This paper has the following highlights. First, a unified theoretical logical framework of "AI-industrial structure optimization-green economic growth" is established, and the industrial structure optimization mechanism of AI affecting green economic growth is revealed from two dimensions of industrial structure upgrading and rationalization. Second, a comprehensive indicator system is used to reflect the whole picture of AI development, which effectively alleviates the problem of variable measurement errors and uses two different instrumental variables to strengthen the control of endogeneity problems. Third, the heterogeneous effects of AI industrial structure optimization effects are captured in terms of factor market distortions, innovative human capital, and the market institutional environment. Fourth, given the adoption of the geographic distance weight matrix and the economic distance weight matrix, the actual effect of AI-enabled green economic growth is further revealed from the perspective of spatial spillover, which provides useful decision-making references for the authorities to comprehensively measure the characteristics of AI development, formulate forward-looking industrial policies and promote inter-regional green synergistic development.#br#