关键词: 人工智能关注, 劳动力技能结构, 就业需求

Abstract: In the context of artificial intelligence (AI) transforming labor markets, this study shifts focus from the downstream effects of AI adoption, such as job displacement, to the earlier stage of corporate strategic thinking. We analyze Chinese A-share listed companies from 2007 to 2023. A company-level measure of AI focus is constructed using text analysis in annual reports, capturing the company's AI focus. We precisely measure employment demand by leveraging large-scale online recruitment data from major Chinese platforms. This allows us to track real-time dynamics in AI-related job postings, including variety, scale, required education, experience, and salary offerings. The research examines how a company's AI focus influences AI-related employment demand.

Empirical findings show that greater AI focus significantly increases both the variety and number of AI job postings, confirming that it creates new demand. This effect depends on a company's internal labor conditions. Companies with a higher proportion of high-skilled labor see stronger job demand creation due to better technological fit. In contrast, companies with inefficient labor allocation use AI focus to correct this imbalance, leading to more employment demand. A key mechanism is that companies rely more on external hiring than on internal training to adapt to technological change. This suggests AI focus may crowd out investment in retraining current staff, thereby generating external demand. The job demand creation effect is stronger in smaller companies, in non-manufacturing industries, and in regions with high-level AI development. Furthermore, AI focus changes the structure of labor demand： it concentrates educational requirements on bachelor's degrees, raises experience requirements, and leads to a balanced increase in salary levels across average, minimum, and maximum wages. This indicates trends of both polarization and inclusive upgrading of job quality.

This study contributes to the literature in three ways. First, it introduces a front-end cognitive perspective by analyzing strategic attention before actual technology adoption, offering early signals of AI's labor market impact. Second, it develops and tests a three-part framework explaining how AI focus affects employment demand by correcting labor misallocation, by leveraging high-skilled labors for better adaptation, and by crowding out internal training investments. This advances beyond simple views of technology replacing jobs. Third, it combines text-based measures of company strategy with detailed, real-time recruitment data, providing strong micro-level evidence on AI-driven labor demand.

The results have important policy implications. First, governments can promote a technology-friendly environment and provide technical interpretation to raise corporate AI awareness, using its job creation potential to ease structural employment pressures. Second, because companies strongly prefer external hiring, governments should guide them to balance recruitment with internal training to avoid over-relying on external human capital. Companies with labor imbalances should optimize their labor structure, while high-skilled companies should use their adaptive advantage. Third, the shift toward bachelor's degrees, greater experience, and balanced wage growth calls for corresponding educational reforms. Governments should guide universities to update AI-related curricula and strengthen continuing education to help labors adapt to new job demands.

Keywords: Artificial Intelligence Focus, Labor Skill Structure, Employment Demand