Utilizing the city-level Artificial Intelligence (AI) patent data and individual data from China Family Panel Studies (CFPS), this paper systematically examines the influence of AI on micro-level labor characteristics in China. The results show that the AI development significantly reduces individuals' likelihood of participating in labor force and decreases the weekly working hours for those employed. AI simultaneously triggers the "labor market exit effect" and the "barrier-to-entry effect". Our analyses also show that the impact of AI development on individuals' working characteristics varies notably across different groups. Specifically, it has a stronger employment replacement effect for older and low-skilled workers, a stronger working-hour reduction effect for younger and high-skilled workers, and a stronger negative effect on both labor force participation and working hours in high-wage cities. Further analyses show that AI development significantly lowers the probability and extent of overwork. These findings showcase both a competitive and collaborative relationship between AI and human workers.