劳动市场条件在 2026 年继续呈现相对平稳。自 2022 年联邦储备为抑制通胀而开始提高利率以来，不稳定性已有缓解，失业率目前已回到去年夏天的水平，且在去年年底联邦政府停摆后略有下降。尽管与伊朗冲突相关的经济和金融冲击持续存在，每周初次申领失业救济人数仍维持在非常低水平，而持续申领失业给付的人数也已不再比前一年同比上升。联邦储备理事会官员 Christopher Waller 指出，劳动力增长疲弱，表示仅需适度新增职位便可使就业大体保持平衡。
AI 焦虑增加了，但其对劳动市场的影响仍是混合性的。Citrini Research 两个月前的情境推演预计 18 个月内可能出现大规模白领职位流失，Anthropic CEO Dario Amodei 最近也警告，科技、金融与咨询等领域的入门职位可能在 1 至 5 年内先被 AI 补强再被取代。Anthropic 也出现同样信号：其 Claude 年化收入在 2025 年底由 90 亿美元上升至 300 亿美元。然而扩张受限，Anthropic 已将高端使用者改为按使用量计费，且随著需求上升，模型表现与可靠性的投诉增加；华尔街日报报导了 AI 计算能力“急剧吃紧”，Goldman Sachs 指出算力已成为扩展 AI 的主要约束。
Wells Fargo 经济学家团队在 Tom Porcelli 领导下认为，AI 压力主要集中在入门白领劳工，而资深劳工机会大体持平或改善。如果以往可作参考，AI 相关失业更可能在衰退中以突发式浪潮出现，而非线性下降；旅游代理在 2001 年衰退及之后才快速下滑，而非 1990 年代网际网路热潮时。由于企业利润仍高，虽然有公司为 AI 花费削减其他成本，但营收仍持续上升，劳动市场短期仍可能保持相对平稳。短期计算能力瓶颈将继续放慢最先进模型的采用，因此本年更像飓风眼前的短暂停歇，而非持续性大规模冲击的开端。
Labor-market conditions are comparatively calm as we head through 2026. Since the Federal Reserve began raising interest rates in 2022 to fight inflation, instability has eased, and the unemployment rate is now back to last summer’s level after edging down since the federal shutdown late last year. Weekly initial jobless claims remain very low despite ongoing economic and financial shocks linked to the Iran conflict, and the number of people continuously receiving unemployment benefits is no longer rising year-over-year. As Fed Governor Christopher Waller noted, labor-force growth is weak enough that even modest hiring can keep employment broadly balanced.
AI anxiety has increased, but its labor impact is still mixed. A Citrini Research scenario two months ago projected massive white-collar losses within 18 months, and Anthropic CEO Dario Amodei recently warned that entry-level roles in technology, finance and consulting could be first augmented then replaced by AI in one to five years. Anthropic’s own trajectory reinforces the concern: its Claude revenue run rate rose from $9 billion to $30 billion by end-2025. Yet scale is constrained—Anthropic shifted power users to usage-based pricing, complaints about model performance and reliability are growing, and the Wall Street Journal reported a sharp AI compute capacity crunch, with Goldman Sachs saying compute has become the binding constraint.
Wells Fargo economists led by Tom Porcelli argue that AI pressure is concentrated in entry-level white-collar jobs while opportunities for experienced workers are mostly unchanged or improved. If the past is a guide, AI-related job destruction tends to arrive in recessionary bursts rather than smooth declines, as shown by travel agents losing many jobs during and after the 2001 recession. With corporate profits still high and revenues still growing, the near-term labor market may remain relatively calm, even as firms scrutinize rising AI spending, tokenmaxxing usage and “AI-washing.” A short-term compute bottleneck may slow broader rollout, so this year could be a brief eye-of-the-hurricane reprieve rather than the start of sustained disruption.