Two things happened to women and AI in the same week, and they point in opposite directions. On Tuesday, LinkedIn published research showing AI roles are among the fastest-growing and best-paid jobs in the economy — median salary of US$177,000 versus US$80,000 for non-AI roles. On Wednesday, the World Economic Forum and LinkedIn jointly warned at the launch of the Global Gender Gap Report 2026 that the gains from that boom are flowing overwhelmingly to men.
“What we found is that the type of roles that are being disrupted by AI tend to be the type of roles that women have been concentrated in,” Saadia Zahidi, a managing board member at the World Economic Forum, told the London launch event. “It is white-collar work that has brought in a lot of women and given them very good livelihoods over the last few decades, and here is exactly those roles that are being disrupted,” added Sue Duke, managing director at LinkedIn.
The numbers behind the warning
The joint report’s topline figure: women hold just one in five AI engineering roles, according to LinkedIn’s Sue Duke. Within AI companies, women are significantly less represented than in comparable non-AI companies — at all levels. Across 27 countries, women hold 13 percent of leadership roles at AI firms, against 20 percent in the broader tech industry.
LinkedIn’s US hiring data fills in the other side of the ledger. Women accounted for 26 percent of hires into jobs requiring AI skills, compared with 50 percent of hires into non-AI positions. Since AI roles carry roughly double the median salary, that split is a direct map of who captures the premium.
The displacement risk is skewed the same way. The International Labour Organization estimates around 29 percent of jobs in female-dominated occupations are exposed to generative AI, versus 16 percent in male-dominated occupations. The Brookings Institution counts 6.1 million US workers in clerical and administrative roles vulnerable to AI, 86 percent of them women.
The other half of the story: the roles are real, and they pay
None of this changes the underlying demand. LinkedIn counted 639,000 AI-related job postings added in the US between 2023 and 2025, and the company’s own numbers show AI postings roughly doubling since 2023. The WEF-LinkedIn framing is not that the jobs are disappearing — it is that a specific group is not getting them. Duke’s argument is that without deliberate investment in bringing women into AI training pathways “from the outset,” the economic upside of the technology could accrue primarily to men.
The report does flag one structural remedy with broader relevance: skills-based hiring rather than credential-based hiring could widen the AI pipeline. LinkedIn’s own data shows more than 95 percent of new hires for head of AI and technical roles hold a bachelor’s degree or higher — a filter that, applied to a labour market where women’s degree attainment outpaces men’s in many countries, is increasingly doing more work than skills testing would.
For career-planners of either gender, the practical read is that the fastest-growing AI job titles — AI engineer, forward deployed engineer, AI consultant — are accessible through demonstrated skills, not only CS degrees. Forward deployed engineers are the breakout hire of 2026, and the role’s definition is closer to consulting and implementation than research science.
New Zealand’s angle on the gap
New Zealand has no AI-specific gender hiring data yet, and nothing in the WEF report is NZ-specific. But the structure the report describes maps closely onto this country’s labour market. Stats NZ’s June quarter data shows women concentrated in the professional, administrative and service occupations that both the ILO and WEF flag as most exposed to generative AI. New Zealand’s public service — one of the country’s largest employers of exactly those white-collar roles — has been running multi-year job reductions that have fallen hardest on administrative and advisory positions. An economy with a 5.6 percent unemployment rate and an AI exposure profile tilted toward female-dominated occupations is the exact scenario the WEF is warning about, and the country’s AI-skills pipeline is small enough that a one-in-five participation rate would be a best case rather than a floor.
The counterweight is that NZ’s tech skills shortage has pushed employers toward skills-first hiring practices for years — long before “skills-based hiring” became a LinkedIn talking point. Whether that practice extends into AI roles, rather than defaulting to degree-gated postings, is the question the next round of local data will need to answer.
What workers can actually do with this
The WEF’s recommendations are aimed at governments and employers, not individuals: invest in women’s AI training pathways, pair them with re-entry funding, redesign work around care. For an individual worker — woman or man — the report’s underlying data suggests three concrete moves. Follow the demand into roles where postings are doubling; treat degree requirements as negotiable where portfolios and certifications can substitute; and watch which adjacent occupations are being “professionalised” by AI rather than automated by it. PwC’s 2026 AI Jobs Barometer found professionalised roles — jobs AI makes more expert-intensive — are growing twice as fast as automated ones, with 42 percent higher wage growth.
The uncomfortable summary is that AI’s labour market is currently rewarding a narrow slice of the workforce while exposing a much broader one. The WEF’s argument is that this is a choice, not a technical inevitability — one that employers and training systems make one hiring decision at a time.
FAQ
What percentage of AI engineering roles do women hold? One in five, according to LinkedIn’s Sue Duke, speaking at the WEF-LinkedIn joint report launch in London on 16 September 2026.
What do AI jobs pay compared to non-AI jobs? LinkedIn’s US data shows a median salary of about US$177,000 for AI roles versus US$80,000 for non-AI roles — roughly double.
Which jobs are most exposed to AI disruption for women? Clerical, administrative, customer service and office-based roles. The ILO estimates 29 percent of jobs in female-dominated occupations are exposed to generative AI, versus 16 percent in male-dominated ones, and Brookings counts 6.1 million vulnerable US workers, 86 percent of them women.
How can hiring become more inclusive in AI? The report points to skills-based hiring over credential filters, early investment in women’s AI training pathways, and continued funding through mid-career transitions.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
