Research report · August 2026

The AI opportunity gap: why women are being left out of the labour market's biggest hiring boom

AI job postings have roughly doubled since 2023 and pay more than double the typical role. Women hold just 26% of the new hires, 20% of Head of AI roles, and 13% of AI C-suite seats worldwide. This report traces the gap from classroom to boardroom using named research from LinkedIn, the World Economic Forum, McKinsey and others — and sets out where training, hiring practice and sponsorship can close it.

CometX Research & Training Centre · Aug 19, 2026 · ~17 min read
Read the findings Jump to sources
Two women collaborating at a laptop in a tech workplace

Photo: #WOCinTechChat, licensed CC BY 2.0

26%

of 2025 U.S. AI hires were women — vs. 50% of non‑AI hires

13%

of AI C‑suite seats held by women, across 27 countries

2.2×

AI pay premium — $177K typical AI role vs. $80K non‑AI

123 yrs

to close the global gender gap at today's pace (WEF)

Executive summary

Artificial intelligence is producing the fastest-growing, best-paid roles in today's labour market — and women are being systematically routed around them. New research from LinkedIn, published August 18, 2026, finds women accounted for just 26% of U.S. AI hires in 2025, with representation falling further at every step up the AI career ladder: 20% of Head of AI roles, and 13% of AI C-suite seats globally.

The gap is not a hiring problem alone. It compounds across a full pipeline — who studies AI-adjacent fields, who is encouraged to use AI at work, who gets sponsored into leadership, and who is protected from (or exposed to) AI-driven job disruption. Closing it is also not a matter of goodwill: the World Economic Forum and LinkedIn's joint research puts the business case plainly.

“Companies that fail to integrate gender parity into AI strategy will miss out on half of the available talent, reducing their capacity for innovation and long-term competitiveness.”

— World Economic Forum & LinkedIn, Gender Parity in the Intelligent Age, March 2025

01 — The AI jobs boom

Some of the best jobs in the economy, growing fast — and going mostly to men

U.S. AI job postings have roughly doubled since 2023, and the roles themselves are shifting quickly — a sign of how central AI has become to hiring plans across sectors, per Axios and LinkedIn's own data.

01

AI job postings have roughly doubled since 2023

02

AI Engineer has overtaken Machine Learning Engineer as LinkedIn's most common AI job title

03

VP of AI postings are up roughly sixfold

04

Forward Deployed Engineer is now the third most common AI occupation in job postings


But the pay premium concentrates in roles women rarely reach. Axios reports that 82% of 2025 hires into “Member of Technical Staff” — a fast-growing research/engineering hybrid role now paying a median $223,000 — went to men. The one AI role that is roughly gender-balanced, data annotation, is also among the lowest-paid.

AI roleTypical listed payShare of hires to women
Data annotator$51,000~50%
Head of AI20%
Director of AI26%
Member of Technical Staff$223,00018%
All AI roles (typical)$177,00026%
All non-AI roles (typical)$80,00050%

Source: LinkedIn, The AI Triple Penalty & The AI Talent Divide (2026); Axios, Aug 18, 2026.

02 — The triple penalty

The gap doesn't happen once. It compounds at three separate points

As LinkedIn's Audrey Lobo-Pulo, Head of Public Policy and Economic Graph for Australia and New Zealand, put it: “AI is creating some of the fastest-growing and highest-value opportunities in the labour market, but women remain underrepresented at every stage: getting hired into AI, reaching its highest-value roles, and leading the companies shaping its future.”

−10pp

The hiring penalty

Women's share of AI roles runs roughly 10 percentage points below their share of non-AI roles.

−5pp

The AI-company penalty

Women's overall workforce share runs about 5 points lower at AI-focused firms than at comparable non-AI firms.

−15pp

The leadership penalty

The gender gap widens by a further 15 points at C-suite level versus roles below the executive team.

Source: LinkedIn, The AI Triple Penalty, Aug 2026.

27 countries surveyed 74 of 75 economies narrowing the AI-skills gap since 2018 Female researcher share ranges 38%–53% across Central Asia

Global picture: Bloomberg, Aug 18, 2026; World Economic Forum, May 2025; Brighterly, STEM statistics 2026.

03 — The leaky pipeline

Representation narrows at every stage, from the classroom to the C-suite

The AI hiring gap sits on top of a much older one. LinkedIn's Economic Graph research finds that for the 2017 cohort of STEM graduates, only 28% of women joined the STEM workforce within a year of graduating, compared with 39% of men — the widest single drop-off in the pipeline. From there, the World Economic Forum's Classroom to C-suite analysis shows women's share of the STEM workforce holding near 29%, before AI hiring and AI leadership data compress it further still.

A woman presenting technical work to colleagues
STEM graduates worldwide 35% Global STEM workforce 29% 2025 U.S. AI hires 26% Head of AI roles 20% AI C-suite (27 countries) 13%

Each bar shows women's share of that population. Figures are drawn from different studies and populations (worldwide STEM graduates, the STEM workforce, U.S. AI hiring, and global AI leadership) — this illustrates the shape of the drop-off, not a single tracked cohort. Sources: edX / UNESCO; World Economic Forum; LinkedIn, 2026.

04 — Why the gap persists

Six mechanisms, each documented separately

No single cause explains the gap. It is produced by ordinary workplace mechanisms that are individually well studied and mutually reinforcing.

Confidence & encouragement

Managers don't push women toward AI tools

Only 21% of entry-level women get management encouragement to use AI, versus 33% of men — and only 37% of entry-level women believe AI will help their careers, per McKinsey & Company data.

Exposure, not just access

Disrupted more often than augmented

In the U.S., 24.1% of men work in roles AI augments versus 20.5% of women — while 33.7% of women work in roles it disrupts, against 25.5% of men, per the joint WEF/LinkedIn white paper.

Underreporting

Women undersell the AI skills they have

LinkedIn's own data suggests women are more likely to underreport AI skills on their profiles than men with equivalent experience, per the World Economic Forum.

Sponsorship, not mentorship

Women get advice; men get advocates

Only 24% of women leaders and 27% of senior women have had a formal mentor, versus 30% and 38% of men — and sponsored employees are 23% more likely to advance, per WomenTech Network and the Centre for Talent Innovation.

The leaky pipeline

The first year after graduation loses the most talent

Only 28% of women STEM graduates joined the STEM workforce a year after graduating, against 39% of men — the single largest drop in the pipeline, per LinkedIn Economic Graph.

The broken rung

The first promotion is where it stalls

For every 100 men promoted to manager, only 93 women are — a gap that compounds every year it goes unaddressed, per McKinsey & Company.

“If women are underrepresented in the roles shaping AI, companies and countries are leaving talent, perspective, and growth potential on the table.”

— LinkedIn, The AI Triple Penalty, 2026

05 — What works
A woman leading a discussion with colleagues around a table

The gap is closing where someone funds the work of closing it

Momentum already exists: the share of LinkedIn members listing AI engineering skills who are women rose from 23.5% in 2018 to 29.4% in early 2025, narrowing in 74 of 75 economies studied, per the World Economic Forum. Structured, funded interventions accelerate it:

  • The WEF's Gender Parity Accelerators are supporting over 1 million women across 18 economies through public-private action plans, engaging 100+ public and 1,500 private partners.
  • EY Ripples and the ITU's AI Skills Accelerator for Girls is upskilling 1,000 young women across 10+ countries through in-person, mentor-led workshops.
  • Microsoft's Women in Digital Business initiative, run with the ILO, equips women entrepreneurs with the digital and business skills an AI-driven economy now demands.
  • Sponsorship consistently outperforms informal mentorship at driving promotion — the lever employers control most directly.
Where CometX fits

A not-for-profit research and training centre, built for exactly this kind of gap

CometX exists to work in the handovers between what research shows and what organizations actually do about it — through applied research, training that produces real deliverables, and equity work measured in who gets hired, promoted, and sponsored. The findings above map directly onto four of our practices.

Training & capacity building

An AI skills pathway, built on a working model

Our workshop, institute, and cohort formats — and Track F's bias-interrupting and inclusive-hiring curriculum — extend directly into a dedicated AI-skills track for women and gender-diverse staff, with a capstone piece of real work at the end, not just notes.

EDIA & women's leadership

Sponsorship cohorts, not mentorship alone

Structured sponsorship pairing women with senior sponsors accountable for advocacy, plus leadership cohorts on negotiation, budget authority, and board readiness — aimed squarely at the 15-point leadership penalty this report documents.

EDIA diagnostics

Measuring where your AI pipeline actually leaks

We audit attraction, selection, progression, pay, experience, and exit by group — the same six points where this report's triple penalty shows up — and hand back a costed action plan with owners and milestones.

Research & knowledge mobilization

Open findings your team can act on

As a not-for-profit, our tools and non-confidential findings publish openly under Creative Commons, with subsidised and pro bono places so smaller organizations and training providers can use them too.


Start a conversation

Most engagements begin with a 30-minute scoping call. Bring the gap you're seeing in your own AI hiring or promotion data — we'll tell you honestly what a realistic scope looks like, fixed-fee, with community rates where they apply.

06 — Recommendations for employers

Six places to look before writing another statement

01

Set hiring targets for AI roles specifically. A "women in tech" goal can hold steady while AI hiring — the fastest-growing, highest-paid segment — quietly reverts to 26/74.

02

Map who is augmented and who is disrupted, by gender, before restructuring. AI transformation plans that skip this step tend to reproduce the exposure gap they should be closing.

03

Fund sponsorship, not just mentoring circles. Sponsorship moves people; mentorship, on its own, mostly moves conversations.

04

Audit the first promotion into technical leadership. The broken rung, not the front door, is where most representation is actually lost.

05

Open AI pathways to non-degree and career-switch talent. Over 91% of AI workers hold a bachelor's degree or higher, rising past 95% in the top-paid roles — a credential filter that quietly filters by gender too.

06

Publish pay by role, not blended averages. The AI premium hides inside job titles; a company-wide gender pay figure can look fine while the AI premium itself skews entirely male.

Sources & methodology

Figures are drawn from the named third-party research below, current as of August 2026 unless otherwise noted. Where studies use different cohorts, countries, or time periods, that is flagged in the text rather than blended into one number. This report was prepared by CometX for informational and business-development purposes and is not a substitute for the underlying studies, which are linked throughout.

6 primary research sources 10 context & analysis sources 6 intervention & training sources