Women and the AI Economy: Opportunities, Barriers and the Skills Gap

Executive summary

Data dashboard

Access & training

The usage gap is closing — the confidence gap isn't, yet

At the population level, the gap in who uses AI at all is smaller than most other digital divides — the OECD puts the overall gender gap in generative-AI use at just 4.2 percentage points, dwarfed by the 53.6-point gap by age. [1] But a Harvard Business School meta-analysis pooling 18 separate studies and 143,008 individuals across 25 countries found women had 22% lower odds of using generative AI than men — a gap that persists once age and occupation are accounted for, and one that widens further in Europe and among older cohorts. [2]

The direction of travel, however, is genuinely encouraging in some markets. In the United States, women's adoption of generative AI tripled over the past year, a faster growth rate than the 2.2x increase seen among men, prompting Deloitte to project adoption parity in the U.S. by the end of 2025. [3] The same pattern of catch-up, though not yet parity, shows up across a dozen additional European countries analyzed by Deloitte UK. [4] Underneath the usage numbers sits a persistent attitude gap: women are consistently more likely than men to expect AI will do more harm than good across personal, professional and public life, and the gender gap in AI use is most pronounced in the 45-and-over age group — where confidence, not access, is the binding constraint. [2]

Employment

The jobs most exposed to automation are disproportionately women's jobs

Adoption gaps matter less than exposure gaps once AI reaches deployment at scale. By one widely-cited estimate, 57% of jobs currently at the highest risk of automation are held by women — a function of occupational segregation into customer service, administrative support and data-entry roles, exactly the task categories generative AI has automated first and most completely. [5] That exposure has already shown up in layoff data: during the major tech layoff waves of 2022–23, women made up around 45% of those let go while comprising only 26–28% of the tech workforce — meaning women were roughly 1.6 times more likely than their workforce share would predict to lose their jobs in those rounds. [6]

This is the sharpest tension in the whole report: the same technology that is helping close the usage gap for women as consumers and workers is, in its current deployment pattern, cutting most deeply into the occupational categories where women are over-represented. Closing the access gap without addressing occupational exposure would only partially protect women's employment prospects.

Entrepreneurship — Canada

Women-led Canadian businesses innovate more and adopt AI less

Canada's 2025 State of Women's Entrepreneurship report finds a real adoption gap at the small-business level: women-owned businesses use AI at a rate of 12.3%, compared to 16.5% among men-owned businesses. [9] That gap sits alongside genuine momentum — the share of Canadian SMEs majority-owned by women has climbed steadily, from 15.6% in 2017 to 17.8% in 2023 to roughly 19% of all businesses in 2024, and self-employed women now make up 37.4% of Canada's self-employed population. [10]

The counter-intuitive finding is on innovation: businesses with majority women ownership (51–99%) post the highest overall innovation rate of any ownership category at 38.0%, ahead of fully women-owned firms (24.7%) and businesses with no women ownership at all (25.6%). [9] Read together, the data suggests the constraint on women entrepreneurs is not appetite for innovation — it is access to the tools, capital and digital-adoption support that would let that innovation translate into AI adoption specifically; only 55% of women-led businesses report having a digital adoption strategy at all. [11]

Leadership

The paradox at the top: women leaders already out-adopt their peers

Every gap documented so far — usage, employment exposure, entrepreneurship — inverts once you reach the executive level. Russell Reynolds Associates' H1 2026 Global Leadership Monitor, surveying 1,172 leaders, found that women in board, CEO, C-level and next-generation leadership roles are on par with, and on several measures ahead of, their male counterparts on generative AI adoption, confidence in their own AI skills, and AI governance practice. [7] Globally, women in the AI workforce overall remain under 30% and drop to roughly 15% in AI leadership specifically — so this leadership-level parity is a genuinely narrow band sitting inside a much larger representation gap, not evidence the gap has closed. [6]

The gap that remains at the top is not capability but perception: a 2025 analysis of how nearly 750 CEOs were portrayed in global media found men CEOs were twice as likely to be described as "innovators" than women CEOs, despite the underlying data showing no capability gap to justify it. [7] Because boards frequently use "innovator" framing as a proxy for future-fit leadership when making succession and promotion decisions, that perception gap functions as an invisible barrier to the CEO pipeline even where the performance data does not support it.

Income

The evidence gap is itself a finding

Unlike access, adoption, employment, entrepreneurship and leadership, there is no single, clean, broadly-cited figure for an "AI income premium" broken out by gender — and that absence is worth stating plainly rather than papering over. What the surrounding evidence implies, without directly measuring it, is a compounding risk: women are more exposed to AI-driven displacement in the occupations where they are concentrated, less represented in the AI-building roles that currently command the largest wage premiums, and less likely to run AI-adopting businesses that capture productivity gains — three separate channels that would each, independently, be expected to widen a gender income gap as AI diffuses further through the economy. Closing this specific measurement gap should be a near-term research priority for national statistical agencies, including Statistics Canada.

At a glance — the gap across the pipeline

Recommendations

Sources

1. OECD, "AI use by individuals surges across the OECD," Jan 2026. oecd.org
2. Stanford Social Innovation Review, "The AI Gender Gap Paradox," Oct 2025, citing Harvard Business School meta-analysis (18 studies, 143,008 individuals, 25 countries). ssir.org
3. Deloitte, "Women and generative AI," Technology, Media & Telecom Predictions 2025. deloitte.com
4. Deloitte UK, "European study on gen AI and trust," cited in Deloitte's 2025 women and generative AI report. deloitte.com
5. SheAI, "Women in AI: Numbers Behind the Gender Gap in Tech," citing World Economic Forum automation-risk data. sheai.co
6. Women & Tech Indiana / WomenTech Network, citing Global Gender Gap Report 2023 and Deloitte. womenandtech.indiana.edu; womentech.net
7. Russell Reynolds Associates, "The AI Gender Gap Isn't Appearing in Leadership Ranks," H1 2026 Global Leadership Monitor. russellreynolds.com
8. Toronto Metropolitan University Diversity Institute / Women Entrepreneurship Knowledge Hub, "State of Women's Entrepreneurship in Canada 2025 Report." torontomu.ca
9. WEKH, "Reaping the Rewards of AI: Transforming Business Responsibly," 2025. wekh.ca
10. She & Success, "How Women Entrepreneurs Can Harness AI and Automation for Technological Growth," citing BDC 2022 data. sheandsuccess.ca