Young Workers Have Mastered AI. The Jobs Are Still the Hard Part
A new World Economic Forum analysis argues the first AI divide — who could access the technology — has largely been crossed, and the next one is sharper: who can turn AI skills into a living. The piece, written by Goodwall CEO Taha Bawa and HP’s global head of social impact programs Michele Malejki and published on 28 September, leans on survey data from young people across more than 50 countries, and it lands on a conclusion that matters directly for New Zealand’s young workers, who face an employment market that has been harder on them than on any other age group.
🔍 THE BOTTOM LINE
Fluency in AI is now nearly universal among young people — but fluency doesn’t pay invoices. The scarce skill of the coming decade isn’t using AI; it’s converting AI skills into work that someone will pay for, and the systems that enable that conversion (internships, mentors, credentials that prove capability) are thin on the ground almost everywhere, including New Zealand.
What the Numbers Say
The Goodwall survey data is striking on the skills side: 91% of surveyed young people use AI at least weekly, 60.2% every day, and among those who assessed their own ability, 97.1% describe themselves as at least somewhat confident using it. This is, by any historical measure, the fastest technology uptake a generation has managed. It sits against a backdrop the ILO quantified in its August youth report: global youth unemployment rose to 12.4% in 2025, with the sharpest rises in higher-income economies.
The opportunity side is where the picture complicates. In the same research, 68.9% of young people identified financial or family pressures as the biggest obstacle between them and their ambitions — for someone expected to contribute to a household, months of unpaid experimentation aren’t affordable. And while 46.8% said they wanted to build their own venture (with another 43.7% open to it), wanting a business and having the capital, network and runway to start one are different things.
The skills themselves also concentrate. As our coverage of LinkedIn’s hiring data noted, the AI roles that do pay well skew heavily toward the young — but they also skew credentialed: LinkedIn found 91% of AI workers hold a bachelor’s degree or higher, above 95% in the best-paid roles. Learning AI, in other words, is easiest for those the economy has already treated well.
Why Skills Alone Don’t Convert
The WEF authors’ core argument is that AI literacy has become a commodity while the things that turn literacy into income remain scarce: real projects to work on, mentors who can name what employers actually value, and credentials that demonstrate capability rather than course completion. Another online course, they argue, adds knowledge without changing economic prospects — a finding that echoes the reskilling-trap analysis we published earlier this year.
There’s a parallel in the employer-side data. When Goodwall and HP’s NextGen AI Alliance — founded with Microsoft and Cognizant — measured its own programme, 967,000 learners had accessed AI skilling opportunities since November 2025. Impressive reach. But the partnership also had to ship 130 physical devices to disconnected young people across 50+ countries, an acknowledgment that access, not skill, was the binding constraint for a slice of learners. Global programmes like Verizon’s free AI training portal face the same test: enrolment is easy to measure, conversion is not.
The ILO’s youth report makes the same point from the institutional side. Its Global Employment Trends for Youth 2026 found global youth unemployment rose to 12.4% in 2025, with 6.1% of jobs held by people aged 15–29 in occupations highly exposed to AI — and the sharpest rises in higher-income economies. ILO employment director Sukti Dasgupta’s summary: “While the direct impact on jobs is still unclear, we must not be complacent and underestimate the risks.”
What This Looks Like in New Zealand
New Zealand sits squarely inside this pattern. The overall unemployment rate was 5.6% in the June 2026 quarter, but youth unemployment ran at 17.2% in the same quarter per Stats NZ data — nearly triple the headline rate. Meanwhile, NZ workers told the Westpac-McDermott Miller survey this month that jobs remain hard to find, with a net 60% agreeing — unchanged from June.
The compounding risk is what career researchers call scarring: young people who spend extended time out of work carry reduced job prospects and lower lifetime wages. That risk grows when the entry-level rungs of a career ladder thin out — which is precisely what the international AI-exposure data has shown, and what the Reserve Bank has flagged for graduate roles.
New Zealand’s counterweights are real but patchy. The Aotearoa AI Summit in Wellington heard this month that 91% of NZ organisations use AI while only 4% have changed how they operate — a gap that is, in the WEF’s framing, a conversion problem wearing a business costume. Employers who redesign workflows create the projects and junior roles where skills become earning; employers who buy licences and stop there don’t.
FAQ
What is the “learning vs earning” divide? It’s the gap between knowing how to use AI (now widespread among young people) and being able to convert that knowledge into a job, income or business — which depends on projects, mentors, networks and credentials rather than skills alone.
How many young people actually use AI? In Goodwall’s 2026 survey data cited by the WEF, 91% use AI at least weekly and 60.2% daily, with 97.1% of self-assessors reporting confidence.
What’s the youth unemployment rate in New Zealand? Stats NZ data put youth unemployment at 17.2% in the June 2026 quarter, against an overall rate of 5.6% — with a net 60% of survey respondents saying jobs are hard to find.
What would actually close the gap? The WEF argument: work-based application (internships, apprenticeships, real projects), mentors who translate skills into pathways, and credentials that prove demonstrated capability. New Zealand’s version is employers who redesign workflows, not just buy AI licences.
🔍 THE BOTTOM LINE
The generation that will work alongside AI has already learned it. Whether that learning becomes earning depends far less on the technology than on the apprenticeship-shaped holes in the labour market — and holes like those get filled by employer decisions, not by more course catalogues.
📰 Sources
- World Economic Forum — The next AI divide is the distance between learning and earning (28 September 2026)
- Goodwall — 2026 Youth & AI Perception Survey; Career Compass research
- ILO — Global Employment Trends for Youth 2026: Back to the future (11 August 2026)
- Stats NZ — Unemployment rate at 5.6 percent, June 2026 quarter; youth unemployment series
- Westpac-McDermott Miller Employment Confidence Index, September quarter (RNZ, 22 September 2026)
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
