How does artificial intelligence investment optimize the skill structure of enterprise labour force? Evidence from Chinese listed companies

Artificial intelligence is becoming a key input in corporate transformation, yet its effects on workforce skill structure remain underexplored. Drawing on skill-biased technological change theory, this study examines AI investment’s effect on the share of highly educated labour and the mediating role of administrative expense ratio. Using panel data of Chinese A-share listed firms (2015–2024), this research finds that a 1% increase in AI investment raises the proportion of employees with a master’s degree or above by about 0.0009% points – significant but modest, reflecting high-skilled talent supply lag. The direct positive effect is partially offset by an indirect negative effect via reduced administrative expense ratio. The effect is strongest in services, weaker in industry, and insignificant in agriculture; it is also significant in non-state-owned enterprises but not in state-owned ones. These findings suggest AI affects skill structure through both production and organizational channels, with implications for firm strategy and policy.