Companies that adopt artificial intelligence are hiring more people, not fewer — but the extra jobs are flowing overwhelmingly to experienced staff, according to a study released on Monday 21 September by researchers at Stanford University and King’s College London. The finding reframes the entry-level problem that has been building all year: AI adoption grows total employment while quietly removing the bottom rung of the ladder.
🔍 THE BOTTOM LINE
Across 41 countries, firms adopting AI raised senior employment 6.7 percent over five years while junior employment fell 3 percent, in research from Bharat Chandar of Stanford and Bouke Klein Teeselink of King’s College London. The gains came from growth at the top, not cuts at the bottom — and the effect deepens in wealthier economies. For a country like New Zealand, where youth unemployment and underutilisation are already climbing, the study describes a mechanism worth watching rather than a verdict already delivered.
What the study found
The paper, “How Does AI Change Labor Demand? Evidence from 41 Countries”, analysed 1.25 billion job postings and 154 million employment records from January 2021 to March 2026. The researchers inferred AI adoption from job advertisements involving generative AI use, then compared hiring patterns at adopting firms against comparable non-adopters.
The results, as Bloomberg reported: senior positions at AI-adopting companies rose 6.7 percent over five years, junior employment declined 3 percent, and the overall share of junior workers fell 1.9 percentage points. Overall hiring increased. Employment in computer and mathematical roles — the most AI-exposed occupations — rose 0.8 percentage points at adopting firms, but even there the centre of gravity shifted senior.
The mechanism the authors describe is asymmetry: “AI is labour-saving for junior workers and labour-expanding for seniors in exposed occupations,” Chandar and Klein Teeselink wrote. The junior-hiring declines showed up across numerous countries including Brazil, Saudi Arabia and the UK, and “run somewhat deeper in richer and more digitised economies.”
How it fits the week’s other data
The study lands as a capstone to an accumulating picture. Stanford’s earlier “Canaries in the Coal Mine” analysis found employment for 22-25 year-olds in AI-exposed US occupations 19 percent below where it would be had it kept pace with less-exposed peers. PwC’s 2026 Global AI Jobs Barometer, analysing over a billion job ads, found the most AI-exposed companies growing headcount faster than the least exposed — 52 percent versus 36 percent — with AI-skilled roles commanding a 62 percent average wage premium. The through-line across all three: AI exposure is not shrinking total employment; it is reshuffling who gets hired and at what level of experience.
Goldman Sachs’ separate report the same week found the impact concentrated in white-collar services — call centres 39 percent below trend in the US, software publishing and consulting shrinking across developed markets — which aligns with the Stanford-KCL finding that the squeeze lands on junior, codified-knowledge work rather than physical labour.
The New Zealand picture
New Zealand’s data shows the same squeeze forming, in a milder form and with different causes. Youth bore the largest increase in underutilisation of any age group in the year to the June quarter, according to Stats NZ, and unemployment overall reached 5.6 percent — the highest in over a decade, with around 166,500 people unemployed. Economists attribute most of that to the post-pandemic slowdown and weak growth rather than to AI.
But the direction of travel for entry-level work is the same as the study’s. SEEK’s August report shows AI skills now required in 4.0 percent of NZ job ads, up 93.3 percent year on year — meaning junior candidates increasingly compete against expectations of AI fluency once reserved for specialists. Goldman’s finding that junior hiring losses deepen in “richer and more digitised economies” places NZ, with its service-heavy economy and rapid AI-skill uptake, closer to the sharper end of the distribution than its size might suggest. There is also a domestic bright spot: roles with low automation exposure — trades, hospitality, hands-on services — are where NZ ad volumes are growing fastest, matching the PwC finding that entry-level jobs demanding “human-intensive” skills like leadership and face-to-face interaction have grown 35 percent since 2019 while other entry roles shrank 10 percent.
FAQ
Are AI-adopting companies cutting jobs? The Stanford and King’s College London study found they added jobs overall — but senior employment rose 6.7 percent while junior employment fell 3 percent over five years. The composition shifted, not just the total.
Why does AI hurt junior roles more than senior ones? The researchers describe AI as labour-saving for junior workers, whose tasks are more codified and routine, and labour-expanding for seniors, who direct and complement it. Junior employment losses were deeper in wealthier, more digitised economies.
What does this mean for NZ graduates? NZ’s youth underutilisation is already the fastest-rising of any age group, though economists mostly attribute that to the economic cycle. The study suggests the entry-level squeeze deepens as economies digitise — which makes AI fluency and human-intensive skills the practical hedge for new entrants.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
