人工智能技术创新与劳动力数字技能溢价

李宏兵　杜佳怡　宫易飞

（北京邮电大学）

摘要：人工智能发展对劳动力市场有长期且深远的影响。在此背景下，本文基于2014—2022年中国家庭追踪调查（CFPS）微观数据，构建并测算数字技能综合评价指标，从个体层面探讨人工智能技术创新对劳动力数字技能溢价的影响。研究发现，人工智能技术创新提升了劳动力数字技能的收入回报，提高了数字技能溢价。机制分析表明，人工智能技术创新通过激发潜在数字人才需求、拉动直接数字人才需求以及优化数字人才就业结构，推升企业对数字技能劳动力的需求，进而提高数字技能溢价。异质性分析表明，人工智能技术创新对数字技能溢价的提高作用在从事非常规任务、资本技术密集型行业以及财政科技支出较高地区表现得更明显。进一步分析发现，人工智能技术创新对数字技能溢价的影响具有明显的收入分布异质性，其提高作用随着收入分位的提高而不断增强，并在高收入群体中表现出更强的技能偏向特征。本文为理解人工智能影响劳动力市场的结构分化提供了微观证据，也为实施差异化、包容性的创新与收入分配政策提供了重要启示。

关键词：人工智能；人力资本；数字技能溢价；工资差距；数字人才需求；技能偏向型

作者简介：李宏兵，北京邮电大学经济管理学院教授、博士生导师，北京，100876；杜佳怡，北京邮电大学经济管理学院硕士研究生；宫易飞，北京邮电大学经济管理学院博士研究生。

基金项目：国家社会科学基金重点项目“数字劳动视域下公共数据高质量开放对稳就业的效应研究”（24AJL016）；北京市宣传文化高层次人才培育资助项目“生成式人工智能赋能北京文化产业数字化发展研究”（京宣传文化人才24-2）；北京邮电大学AI4S探索计划项目（2025AI4S11）

引用格式：李宏兵，杜佳怡，宫易飞. 人工智能技术创新与劳动力数字技能溢价[J]. 经济与管理研究，2026，47（8）：56-70.

AI Technological Innovation and Digital Skill Premium of Labor

LI Hongbing, DU Jiayi, GONG Yifei

(Beijing University of Posts and Telecommunications, Beijing 100876)

Abstract: Artificial intelligence (AI) has become a key driver of labor market transformation, profoundly reshaping skill demand and wage distribution. As AI technologies continue to spread across industries, digital skills have emerged as an increasingly important component of human capital and a major source of labor income. Using nationally representative microdata from the China Family Panel Studies (CFPS) from 2014 to 2022, this paper constructs a comprehensive digital skill index based on four dimensions—digital access, digital technology usage, Internet application skills, and the importance of digital information channels—and examines the impact of AI technological innovation on digital skill premium of labor.

The results show that AI technological innovation increases the income returns to digital skills, thereby enhancing the digital skill premium. The findings remain robust after endogeneity analysis and robustness tests. Mechanism analysis indicates that AI promotes the digital skill premium by stimulating potential demand for digital talent, expanding direct demand for digital talent, and optimizing the employment structure of digital talent. Heterogeneity analysis further reveals that the positive effect is more pronounced among workers performing non-routine tasks, those employed in capital- and technology-intensive industries, and those located in regions with relatively abundant labor resources. Moreover, the recentered influence function (RIF) quantile regression shows significant distributional heterogeneity, with stronger effects at higher income quantiles and a more pronounced skill bias at the upper end of the income distribution. Therefore, efforts should be made to actively promote a sci-tech innovation-oriented approach to drive a reasonable increase in labor income, establish a multi-tiered differentiated system to enhance digital skills, and simultaneously strengthen the systematic accumulation of human capital. Meanwhile, focus should be placed on planning the regional distribution of technological development in a reasonable way and refining mechanisms for regulating income distribution.

The marginal contributions are threefold. First, unlike conventional skill measures based on educational attainment or task characteristics, this paper develops a multidimensional measure of digital skills that better reflects the human capital requirements of the digital economy. Second, it proposes and empirically verifies a three-stage mechanism linking AI technological innovation to the digital skill premium: stimulating potential demand for digital talent, expanding direct demand for digital talent, and optimizing the employment structure of digital talent, thereby providing new evidence on how AI reshapes labor market outcomes. Third, by employing the RIF quantile regression, it moves beyond average treatment effects to uncover the distributional consequences of AI. The findings enrich the literature on the labor market effects of AI and provide policy implications for promoting inclusive digital transformation and improving the allocation of digital human capital.

Keywords: artificial intelligence; human capital; digital skill premium; wage gap; demand for digital talent; skill-biased

下载全文