跨职业差异是主要分化来源。在白领岗位中，诸如 lawyers、accountants、software developers 等高知识职位，在初级与高级层级都显示出较高 AI 使用率，但在同一产业的低薪岗位明显落后。Nobel 奖得主 Daron Acemoglu 与 Chris Pissarides 指出，AI 与更高的抽象与量化能力具有互补性：技术愈先进，人类智慧对结果的影响愈大。因此 AI 有机会提升高端劳动者生产力，进而扩大利润或工资差距，利益集中在原本已具资源者。
性别差距同样持续：在不同产业中，women 使用 AI 的可能性比 men 低 20%。Google 首席经济学家 Fabien Curto Millet 指出，差距可透过培训缩小；其团队 2025 年研究发现，对英国 workers 的一项训练让55岁以上 women 的每日使用率增加了三倍。这与 FT 发现一致：企业培训是推动职场 AI 使用的最大单一驱动力。AI 最重度使用者反而是 30 岁且任职年资较长者而非最年轻员工，支援 OpenAI 经济学家 Ronni Chatterji 的观察：AI 有助成熟专家提升效率。这加剧了职涯金字塔底部受损风险，即高阶工作者可将原本由 junior 承担的工作交由 AI，限制新进员工的技能累积；若教育制度未将深度专长作为激励主轴，差距可能持续扩大。
An FT-Focaldata poll of 4,000 workers in the US and UK shows AI adoption is highly unequal: over 60% of higher-paid workers use AI daily, compared with only 16% of lower-paid workers. The UK fieldwork covered 2,365 workers (26 February–2 March 2026), and the US fieldwork covered 1,754 workers (6–9 March 2026), with data weighted to be nationally representative by age, gender, region, education, ethnicity, political interest, and past voting behavior, indicating usage rises with pay and cognitive resources rather than spreading evenly across roles.
Cross-occupation variation is the major divider. In white-collar jobs, high-knowledge roles such as lawyers, accountants, and software developers show high AI use at both junior and senior levels, while lower-paid roles in the same industries lag strongly. Nobel laureates Daron Acemoglu and Chris Pissarides argue this reflects complementarity between AI and higher abstract/quantitative ability: as technology gets more advanced, human capability matters more. AI is therefore likely to increase top-end productivity, potentially widening wage or capital share gaps as gains flow mainly to already advantaged workers.
The gender divide remains: women are about 20% less likely to use AI than men across sectors. Google chief economist Fabien Curto Millet says training can narrow this gap; his team’s 2025 study found a training program for UK workers tripled daily AI use among women over 55. This matches FT data that corporate training is the strongest single driver of workplace AI use. The heaviest users were also those in their thirties with longer tenure, not the youngest workers, supporting OpenAI economist Ronni Chatterji’s view that AI amplifies established expertise. That pattern raises concern that AI may hollow out the bottom of the career pyramid, as junior tasks shift upward to AI under senior staff, limiting early-career skill formation and potentially worsening long-run inequality.