AI will decide who gets hired, who gets a loan, who gets a diagnosis. The people building it don't reflect the world it's supposed to serve. Here's what the data says.
Women earn 35% of STEM degrees. That sounds like progress. But follow them through their careers and the number drops at every stage. By the workforce, it's 28%. By management, 24%. By AI research? 12%.
A 23-point drop from graduation to leadership. Two-thirds of the pipeline leaks before reaching the rooms where AI gets built.
From graduation to the C-suite, women's representation drops 23 percentage points.
Each square = 1 AI researcher out of 100
Sources: UNESCO · U.S. Census Bureau · Stanford HAI AI Index 2021-2026 · WEF/LinkedIn 2025
Women don't just face underrepresentation in AI. They face overexposure to what AI displaces.
ILO studied GenAI's impact across the global workforce. In high-income countries, 9.6% of women's jobs are at high risk of automation. For men, it's 3.5%. Nearly three times the exposure. That gap holds in 88% of countries studied.
Women are concentrated in clerical and administrative roles that GenAI can replicate. The new AI engineering roles being created? Those skew 71% male.
Sources: ILO-NASK, "Generative AI and Jobs: 2025 Update" · WEF Future of Jobs Report 2025
Here's what surprised me most. Lean In surveyed 1,000 U.S. adults in March 2026. Women and men use AI at work at nearly identical rates: 73% vs 78%.
But managers encourage men to use AI 23% more often. And men are 50% more likely to be praised for using it. Same behavior. Different recognition.
Meanwhile, men adopted GenAI tools faster overall: 44% vs 33% of US professionals in 2024. But women's adoption tripled year-over-year, outpacing men's 2.2x growth.
Sources: Lean In / Wired Research, March 2026 · Deloitte, "Women and Generative AI," 2024
The gap varies by geography. Singapore leads: 35% of its AI talent is women. Saudi Arabia follows at 32%. Germany, despite being an economic powerhouse, sits at 20%.
And in Europe, the picture is getting worse. McKinsey found the share of women in core tech roles dropped from 22% to 19% between 2023 and 2025.
Sources: OECD.AI/LinkedIn · McKinsey, "Women in Tech and AI in Europe," 2025
economies saw their AI gender gap narrow between 2018 and 2025.
Women's share of AI skills on LinkedIn grew from 23.5% to 29.4% in seven years. The question isn't whether the gap can close. It's whether it closes fast enough. The AI systems being built right now will shape hiring, healthcare, lending, and law enforcement for the next decade.
88% of AI researchers are men. That's not a statistic. That's a design decision made by default.
Sources: WEF/LinkedIn 2025 · Deloitte 2024 · Stanford HAI