在招聘的 AI 时代，人才招聘人员反映，以前以人为本且令人满意的配对任务正变得更困难。Greenhouse 的 Daniel Chait 将其称为「AI doom loop」：用于制作履历与优化求职信的便捷工具提高了申请量，但企业却难以将高品质候选人与 AI 生成的垃圾内容区分开来。Greenhouse 的研究显示，自 2022 年以来，每个职位的申请量已翻倍以上，招募者面临「信号越来越少」的问题。
然而，多数招聘人员仍认为 AI 有营运效益。Korn Ferry 的 Bryan Ackermann 表示，自动化是为了「为那些人类时刻清理场地」，而 Cisco 的 Scott McGuckin 则认为 AI 带来「转型效应」，可让团队把时间留给更具战略与创意的工作。生成式 AI 被用来起草职位描述与评估，BCG 调查显示许多组织在招聘初始阶段就使用 AI；它也可建立大量候选名单与短名单，并加快筛选：McLaren 在9月收到超过 21,000 份毕业生计划申请，故导入 Microsoft 工具做初筛；Eightfold 推出 AI Interviewer 支援技术面试，Randstad 也启用 AI 聊天机器人进行筛选，印度专用机器人可使用 30 多种当地语言沟通。
法律与伦理限制仍然存在。欧盟 AI 法规禁止将 AI 用于作出招聘决策，且许多人认为人类判断是根本价值所在。Robert Half 的 Matt Weston 认为，没有 AI 模型能辨识个性，只有把技能和资历与人的特质结合才能完成真正成功的配置。确实，高层招聘（较少受量化压力）仍属以人为主的流程，Finatal 首席执行官 Paul Blant 称其为约 95% 的人类判断；Randstad 的 Martin de Weerdt 则认为 AI 应用于「再个人化」招聘，使招聘方能在原本人力不足的规模下与候选人建立有意义对话。Greenhouse 的 Chait 进一步预测长期而言，招聘将被「全面重想」，未来候选人可由了解其经历与下一步意愿的求职代理人代表，而企业端则由招聘主管代理人回应关键条件与职缺匹配，双方都以人为核心互诉其事。
In the AI era of hiring, recruiters report that what used to be a satisfying human match-making task is becoming harder. Daniel Chait of Greenhouse calls the cycle an “AI doom loop”: easy tools for résumé building and cover-letter optimization boost application volume, but firms cannot distinguish high-quality candidates from AI-generated filler. Greenhouse research shows applications per job have more than doubled since 2022, leaving recruiters with “less and less signal.”
Yet most recruiters still see operational benefits. Korn Ferry’s Bryan Ackermann says automation is meant to “clear the decks for those human moments,” while Cisco’s Scott McGuckin says AI has a “transformative effect” by freeing staff for strategic, creative work. Generative AI is used to draft job descriptions and assessments, and BCG reports many organisations use AI from the hiring start; it also builds wide longlists and shortlists and speeds filtering: McLaren received more than 21,000 graduate-scheme applications and adopted Microsoft tools for initial screening; Eightfold launched AI Interviewer for technical interviews, and Randstad uses screening chatbots, including one in India that can communicate in 30-plus local languages.
Legal and ethical limits remain. The EU AI Act bars use of AI for hiring decisions, and many practitioners insist human judgement is essential: Matt Weston says no AI model can detect personality, and true placement requires combining skills and credentials with the human element. Executive recruitment remains largely human, with Paul Blant, chief executive of Finatal, calling it about 95% human judgement. Randstad’s Martin de Weerdt says AI should “re-personalise” recruitment, enabling scale and meaningful conversation; Greenhouse’s Chait predicts a long-term reinvention where candidates are represented by job-seeking agents who know their experience and next interests, and hiring-manager agents match openings to criteria, with people at the center.