Women Are Avoiding the Very Technology That Threatens Their Jobs Most. An Expert Just Warned of a 'Two-Tiered AI Economy.' Here Is What the Data Shows.

Fortune ran a headline this week that deserves more attention than it has received: 'Women are avoiding the very technology that threatens them most, as expert warns of a two-tiered AI economy approaching.' The expert quoted is Dr. Sarah Kimani, a labor economist at Georgetown's McCourt School of Public Policy, who has been tracking AI adoption patterns by gender across a panel of 50,000 US workers since 2023. Her finding: men are adopting AI tools for professional work at approximately double the rate of women. The occupations with the highest rate of female employment — administrative assistants, paralegals, medical records specialists, customer service representatives — are simultaneously the occupations with the highest AI exposure. The workers most at risk from AI are the least likely to be using AI to protect themselves. This is one of the most important labor market dynamics in 2026, and it is almost entirely absent from the mainstream AI discourse.

The Data: Where the Gender AI Gap Is Widest

• Professional AI tool adoption: a 2025 McKinsey survey found that 67% of men in knowledge economy roles used AI tools for professional work weekly, versus 41% of women in equivalent roles. The gap is wider in industries with higher AI capability — technology, finance, legal — and narrower in industries with less AI penetration.
• AI in male-dominated vs female-dominated occupations: men's highest-employment occupations (software development, management, finance) have high AI tool adoption and also have strong leverage to use AI for productivity amplification rather than replacement. Women's highest-employment occupations (administrative support, healthcare support, education) have high AI exposure for task automation — meaning the AI risk is higher and the AI-as-tool adoption is lower.
• The education correlation: AI tool adoption is higher among workers with four-year degrees. Women earn 57% of bachelor's degrees and 59% of master's degrees — educational attainment is not the explanation. The gap appears to correlate with workplace culture, social permission to experiment with new tools, and the specific nature of AI tools available (which were initially built primarily by and for technical roles).

Why Women Are Less Likely to Adopt AI Tools

The reasons are not simple and the evidence does not support any single explanation. Several factors appear to be contributing simultaneously.

• Risk-aversion in highly scrutinized roles: research consistently shows women in professional settings face higher scrutiny for errors than their male counterparts. In this context, using an AI tool that might produce an error — and having that error attributed to you — may represent a disproportionate career risk for women. Men, who face less career penalty for equivalent errors, may be more willing to experiment.
• Workplace culture and social permission: in many organizations, adopting new technology tools requires unspoken social permission — the tacit approval of colleagues and managers that experimentation is acceptable. Women in male-dominated workplaces may perceive less of this social permission to experiment, leading to lower adoption even when the tools are available.
• AI tool design biases: early large language models were primarily trained on text from the internet, which skews toward male-dominated technical communities. Tools designed primarily by men, for initially technical use cases, may feel less intuitive or relevant to workers in caregiving, administrative, or healthcare support roles.
• The relevance gap: the most heavily marketed AI tools — code generation, technical writing, data analysis — may simply not be the right tools for the specific tasks that dominate female-majority occupations. Healthcare documentation AI, administrative scheduling AI, and education support AI are less prominent in the AI discourse despite being more directly relevant.

What the Two-Tiered AI Economy Actually Looks Like

Dr. Kimani's 'two-tiered AI economy' concept describes a scenario where AI adoption patterns along gender lines produce two divergent economic outcomes. Workers who adopt AI for professional use — amplifying their productivity and thereby increasing their economic value — see rising wages and job security. Workers who do not adopt AI — whether because they lack access, face cultural barriers, or are in occupations where appropriate AI tools do not yet exist — face stagnant productivity and increasing displacement pressure from AI-automated processes. If the gender AI adoption gap persists, the economic divergence compounds over time.