统计研究

在明确界定人工智能与实体经济产业范围的基础上，以2013—2023年我国公开发明授权专利数据为样本，结合专利分类号与产业的对应关系，采用专利共分类方法，系统测算我国人工智能与实体经济融合水平，并从时间、产业、空间、网络4个维度剖析融合特征。研究发现，在时间维度上，2013—2023年我国人工智能与实体经济融合水平呈不断上升趋势；在产业维度上，融合水平呈现第三、第二、第一产业依次递减的特征，人工智能基础层对融合的支撑作用持续强化；在地区维度上，融合水平存在明显的空间分化特征，即东部、中部、西部和东北四大地区呈现逐级递减的梯度分布格局，且各地区内部也逐渐分化形成多梯度格局，京津冀、长三角、珠三角和成渝四大城市群内部核心城市与非核心城市的融合水平梯度分化明显；在网络维度上，融合的广度和深度均有增加，但仍有较大提升空间，人工智能产业在融合网络中的影响范围扩展速度快于实体经济产业，融合网络逐渐形成以人工智能产业为中心、实体经济产业分布于外围的“核心–边缘”结构特征。研究结论为纵深推进“人工智能+”行动、打造智能经济新形态提供参考。

Based on the clearly defined industrial scopes of artificial intelligence(AI) and the real economy, this paper takes China's authorized invention patents publicly issued from 2013 to 2023 as research samples. By matching patent classification numbers with corresponding industrial categories, the patent co-classification method is adopted to systematically measure the integration level between AI and the real economy in China, and the integration characteristics are analyzed from four dimensions: time, industry, region and network. The research conclusions are as follows. First, from the temporal dimension, the integration level of AI and the real economy in China maintained a sustained upward trend during 2013—2023. Second, from the industrial dimension, the integration level decreases sequentially among the tertiary, secondary and primary industries, and the basic layer of AI exerts an increasingly prominent supporting effect on industrial integration. Third, from the regional dimension, prominent spatial differentiation exists in the integration level. The integration degree presents a descending gradient pattern across the Eastern, Central, Western and Northeast regions of China, and multi-gradient differentiation has gradually taken shape within each region. Obvious gradient gaps in integration levels can also be observed between core and non-core cities inside the four major urban agglomerations, namely the Beijing-Tianjin-Hebei, Yangtze River Delta, Pearl River Delta and Chengdu-Chongqing urban agglomerations. Fourth, from the network dimension, both the breadth and depth of integration have expanded, while there is still considerable room for further improvement. The influence scope of AI industries expands faster than that of real economy industries, and the integration network has gradually evolved into a core-periphery structure with AI industries at the center and real economy industries distributed in the peripheral areas. The research conclusions offer empirical references for further promoting the “AI+” initiative and fostering new forms of intelligent economy.