在截至2025年5月的12个月里，降幅最大的是客户服务代表，就业人数减少130,180人，或4.8%。不包括医疗、法律和行政秘书与助理在内的秘书和助理人数减少31,030人，或1.8%；批发和制造业销售代表（不含技术和科学产品）减少28,670人，或2.3%。自2022年5月以来，也就是ChatGPT在那年稍后推出之前，18个职业中降幅最剧烈的是信用授权员、审核员和文员，下降26.2%；广播播音员和电台DJ，下降20.8%；以及销售工程师，下降13.2%。
这些数据表明，AI可能已经在重塑招聘模式。Goldman Sachs 经济学家表示，高度受影响的职业其职缺已降至疫情前水平以下，而AI增强型职位或受影响较小的职位则下降得更为缓慢。BLS在其2024年11月的报告中提醒，其列出的18个职业并非穷尽或定论，而只是目前合理预期会受到AI驱动影响的例子。即便如此，受影响职位的连续下降，与总就业成长0.8%形成对比，提供了一个早期的量化讯号：AI部署可能正在促成劳动市场的重新配置，而非这些职位的全面扩张。
Several US occupations widely viewed as exposed to artificial intelligence saw another year of heavy job losses in 2025, with customer service representatives, certain secretaries, and some sales roles among the weakest. The Bureau of Labor Statistics identified 18 AI-exposed occupations, totaling about 10 million jobs, and said employment in that group fell 0.2% between May 2024 and May 2025, while overall US employment rose 0.8% over the same period. Excluding medical secretaries and assistants, a fast-growing healthcare-linked category, employment across the other 17 occupations declined 1.6% for the second straight year.
The largest declines included customer service representatives, whose employment fell by 130,180, or 4.8%, over the 12 months through May 2025. Secretaries and assistants except medical, legal, and executive decreased by 31,030, or 1.8%, and wholesale and manufacturing sales representatives excluding those for technical and scientific products dropped by 28,670, or 2.3%. Since May 2022, before ChatGPT was introduced later that year, the steepest declines among the 18 occupations were in credit authorizers, checkers, and clerks, down 26.2%, broadcast announcers and radio disc jockeys, down 20.8%, and sales engineers, down 13.2%.
The data suggest AI may already be reshaping hiring patterns, with Goldman Sachs economists saying highly exposed occupations have seen job openings fall below pre-pandemic levels, while AI-augmentation roles or less exposed roles have declined more gradually. The BLS cautioned in its November 2024 report that its list of 18 occupations was not exhaustive or definitive, but instead represented examples where a reasonable expectation of AI-driven impact currently exists. Even so, the repeated declines in exposed jobs, contrasted with 0.8% total employment growth, provide an early quantitative signal that AI deployment may be contributing to labor-market reallocation rather than broad-based expansion in these roles.