Research

Australia's Central Bank Finds AI Hits Young Workers First

The Reserve Bank of Australia finds young workers in AI-exposed roles like software faring worst since ChatGPT's launch — but causation remains unproven, and macro factors may explain much of it.

Australia's central bank finds AI hitting young workers first, especially in software and customer service
Australia's central bank finds AI hitting young workers first, especially in software and customer serviceSammy-Sander / pixabay
By James Calloway5 min read

Why it matters

  • RBA analysis: job postings fell across all groups since ChatGPT's late-2022 launch, steepest in the highest AI-exposure group.
  • Stanford (Aug 2025): employment of 22-25-year-olds in AI-exposed US occupations down 13%, nearly 20% for young software developers.
  • Anthropic internal review: junior share of technical hires fell from 27% to 9% since early 2025; AI agents log 6x human hours as of August 2026.
  • Denmark study of 25,000 workers: 64% AI adoption but no wage or hours changes; average time saved just 2.8%.
  • WEF projects a net gain of 78 million jobs by 2030; LinkedIn AI job postings grew sixfold in one year.

Australia's central bank has found that AI has barely displaced any jobs so far — but young workers entering highly AI-exposed fields like software development are faring measurably worse than everyone else.

The analysis, authored by Mike Major for the Reserve Bank of Australia (RBA), deliberately did not hunt for economy-wide disruption. Instead, Major built a monitoring system aimed at the places where effects should appear first: entry-level workers and occupations with high AI exposure, such as programming, clerical work, and customer service.

The stakes are straightforward. If generative AI reshapes labor markets, central banks need early signals for wage growth, inflation, and employment policy — and entry-level hiring is where those signals should show up before anywhere else.

What did the RBA actually measure?

According to the analysis, job postings have fallen across all groups since ChatGPT launched in late 2022. The steepest drop occurred in the highest-exposure group.

Within that group, workers aged 15 to 24 and young college graduates show weaker employment rates and higher unemployment than older workers doing comparable work.

Major is explicit about the limits of his own findings. The study does not establish direct causation. It measures theoretical exposure to AI, not actual use of AI by companies. Post-pandemic shifts in the business cycle, or changes in remote work patterns, could also be driving the numbers.

Hiring data from SEEK, one of Australia's largest job platforms, complicates the displacement story further: according to that data, AI is complementing existing work while also creating new demand for AI skills.

Takeaway: the RBA sees a real pattern among young workers, but calls it a pattern — not proof that AI took their jobs.

Do other studies show the same thing?

Yes, with caveats attached to nearly every one.

A Stanford study from August 2025, based on millions of US payroll records, found a similar pattern. Employment among 22- to 25-year-olds in AI-exposed occupations fell 13 percent in relative terms since late 2022. Among young software developers, the decline was nearly 20 percent. More experienced workers in the same occupations held steady or gained ground.

A March 2026 study by Anthropic, the maker of the Claude AI models, also found signs of declining hiring among 22- to 25-year-olds in exposed occupations, with job-finding rates down about 14 percent. It found no systematic rise in overall unemployment, however. The study's authors stress that theoretical automation potential and actual AI use remain far apart.

Anthropic also offered a company-level example — and it is stark. According to an internal review, the share of junior hires in technical roles has fallen from 27 percent to 9 percent since early 2025. In August 2026, AI agents across the company were already logging six times as many hours as human employees.

Anthropic concedes that, as an extremely AI-intensive software company, it is a special case. Even so, its researchers warn of what they call a "tragedy of the cognitive commons" — the risk that disappearing junior roles lead to a future shortage of experienced professionals, assuming those professionals are still needed at all.

Takeaway: three independent datasets now point the same direction for workers in their early twenties, though none of them proves AI is the sole cause.

Is there counter-evidence?

Plenty.

A large study of 25,000 workers in Denmark found no economy-wide impact at all. Despite a 64 percent AI adoption rate, researchers found no measurable changes in wages or hours worked. Workers saved an average of just 2.8 percent of their working time — a modest figure for a technology routinely described as transformative.

A January 2026 analysis of 10.6 million LinkedIn profiles and US unemployment data pushes back even harder on the AI explanation. The research team, led by Morgan Frank at the University of Pittsburgh, found that the problems in exposed occupations started in early 2022 — months before ChatGPT launched. Macroeconomic factors, such as the Federal Reserve's shift in interest rate policy or a correction after pandemic-era hiring surges, may play a bigger role than previously assumed.

If the downturn predates the technology, attributing it to the technology becomes much harder.

Takeaway: the strongest contrary evidence suggests macroeconomics, not AI, may explain much of the weak entry-level market.

Where is AI creating jobs?

At the same time, AI is generating new demand for labor rather than only destroying it.

According to the World Economic Forum's "Future of Jobs Report 2025", 77 percent of surveyed companies are prioritizing upskilling over job cuts. WEF researchers project a net gain of 78 million jobs by 2030.

LinkedIn data points in the same direction: AI-related job postings grew sixfold in one year.

These figures come with their own caveat. Projected net job gains for 2030 are forecasts, not observations, and the WEF's numbers reflect surveyed company intentions rather than measured hiring outcomes.

Takeaway: demand for AI skills is rising fast, but most of the optimistic numbers are projections, not payroll data.

Why this matters now

The RBA analysis lands in the middle of an unresolved argument among economists. Young workers in AI-exposed occupations — programming, clerical work, customer service — are consistently underperforming their older colleagues across US, Australian, and company-level data. Yet the Denmark study shows no wage or hours effects at 64 percent adoption, and the Pittsburgh team shows the trouble began before ChatGPT existed.

For central bankers, the practical question is which signal to trust. A labor market shock that starts with 22- to-25-year-olds would eventually feed into wage bargaining, consumption, and inflation expectations. A cyclical correction in tech hiring, by contrast, would fade on its own.

Major's monitoring system is built for exactly this ambiguity: it watches the groups where AI effects should surface first, without claiming they already have. The next rounds of data — particularly on whether junior hiring recovers as AI adoption matures — will determine whether Anthropic's "tragedy of the cognitive commons" becomes a policy problem or a footnote.

Original: the-decoder.de

Share this article:

More from James Calloway

James Calloway

Show full bio

News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.

604 articles

Related articles

  1. AI Was Supposed to Hit New Grads Hard. Unemployment Data Says Otherwise
  2. IMF's Georgieva Warns AI Is Hitting Global Jobs Like a Tsunami
  3. 65% of Australian Frontline Workers Use AI — 46% Got No Rules
  4. 42% of New College Grads Are Underemployed as AI Rewrites Hiring
  5. AI Adoption Cut Junior Hiring by 9%. Aviation Already Knows What Comes Next.

« Previous article