本报告为高盛全球经济分析专题，研究人工智能对劳动力替代的幅度与节奏。作者指出，近期市场对AI即刻引发大规模失业的担忧已有所降温，但劳动力市场数据开始显露与AI相关的招聘逆风，讨论焦点转向AI带来的劳动力成本节约能否支撑持续攀升的AI资本开支。报告采用三种实证路径衡量历史上技术对劳动力的替代：总量层面的技术冲击显示，技术进步带来的生产率提升会暂时推高失业率，随后逐步消退，且主要源于离职裁员增加而非招聘放缓；行业层面的生产率冲击效果与之相近，但裁员反应在生产率大幅提升时更为剧烈；对曾受技术显著冲击的职业考察则显示，当技术使劳动力变得可替代时，相关岗位的就业损失更为明显。据此报告得出三点结论：在生产率提升的基准情形下，相当比例劳动者在AI转型期可能被暂时替代；就业减少更多由裁员而非招聘放缓驱动，因此AI相关裁员公告可作为较好的实时观察指标；行业与职业层面分析意味着劳动力替代存在上行风险。
Concerns of imminent AI-related job losses have cooled. But recent labor market data show some signs of AI-related hiring headwinds, and market participants are increasingly focused on whether AI labor cost savings will be sufficient to justify investment. In this Global Economics Analyst, we consider three empirical approaches to quantify how technology has historically displaced labor and the implications for AI labor disruption. First, aggregate shocks imply that each 1% increase in labor productivity due to technological innovation leads to a temporary 0.3-0.4pp peak boost to the unemployment rate that fades after two years, driven by a cumulative increase in gross job losses and a smaller drag on gross hiring. Second, industry-level productivity shocks imply very similar effects on unemployment and gross job gains and losses, but larger impacts on job losses in the late 1990s and following larger productivity shocks. Third, examining 25 occupations exposed to significant technolog