
Are Young Australians Feeling AI’s Job Impact First?

Young Australians in occupations whose tasks are more exposed to AI show tentative signs of a weaker job market, although Australia has seen little broad AI-driven employment disruption. On 9 November 2026, Reserve Bank staff analysis by Mike Major identified slightly weaker job-advertising flows in exposed occupations and softer employment outcomes for younger people working in them. The analysis describes associations in Australian data; it cannot establish that AI caused a lost job or an unfilled vacancy.
In her speech at the UBS Australasia Conference in Sydney on 9 November 2026, RBA Assistant Governor Sarah Hunter said, “Even where AI replaces tasks, it may not replace the whole job.” A chart of the bank’s discussions with firms showed that roughly three-quarters reported no AI-related change in headcount over the previous year. About 12% reported slower hiring or leaving roles unfilled, compared with about 8% that reduced headcount and 6% that increased it.
Employment data show resilience across exposed occupations
Employment has grown in every occupation group ranked by AI exposure since the analysis’s late-2022 benchmark. Growth in employment and total hours worked has been somewhat weaker in the most exposed groups, but the difference is modest. Restricting the comparison to market-sector jobs produces a similar pattern, so stronger hiring in government-funded services alone does not explain the gap.
Worker flows are quieter still. Job-finding rates for unemployed people and movements from employment into unemployment show little sustained difference across exposure groups. A large wave of AI-related dismissals would be expected to leave a clearer mark on the latter measure. The unemployment rate by previous occupation also needs care: it can rise when more people seek work, even without a corresponding increase in layoffs.
The broad figures have not turned sharply against exposed occupations. The clearest signal appears only when the analysis separates workers by both age and the tasks in their jobs.
Firms report more restraint in hiring than cuts

The firm reports locate a different possible point of adjustment: the decision to create or fill a role. Slower recruitment can reduce openings for someone entering an occupation while leaving existing staff employed. That mechanism explains how a young applicant might encounter a tighter market before a change becomes visible in the workforce total.
The liaison figures are firms’ own accounts of whether AI affected staffing, rather than a tally of Australian jobs created or destroyed by the technology. They also include businesses that added workers because of AI, so the responses do not describe one uniform strategy. Still, the reported balance matters: leaving openings vacant was more common than reducing existing headcount among firms that reported an effect.
Most firms in the liaison group did not expect AI to change their headcount over the following year. That expectation concerns the firms consulted and their current plans; it does not measure how many new jobs they will advertise or what skills future recruits will need.
Job advertisements reveal a narrow hiring slowdown
Vacancy data offer a more direct look at demand before a person is hired. The occupation analysis maps Jobs and Skills Australia postings and SEEK advertisements into exposure groups. In both series, new advertisements in the most AI-exposed occupations have slowed slightly more than those in other groups. Within the highest-exposure group, general clerks, software and applications programmers, ICT business and systems analysts, call or contact centre workers, and receptionists contributed most to the decline in new Jobs and Skills Australia advertisements from the late-2022 benchmark.
The distinction is between a decline in the flow of new advertisements and a decline in the number of people already working. Advertisement trends across exposure groups have generally followed the same cycle, with only a small gap between them. The flow can also change as employers adjust recruitment budgets or use different channels, so the comparison points to a possible soft spot rather than a measured number of jobs replaced by AI.
There is a second signal in the wording of vacancies: a growing share mention AI-related skills. That may show changing demand for workers who can use the technology, but some employers may simply be updating the language of their listings. It cannot by itself tell whether another task or position has disappeared.
The clearest divergence is among younger workers

Age and occupation together reveal the sharpest contrast. In the most AI-exposed occupations, younger workers have had weaker employment outcomes and larger increases in unemployment than older workers; the same relationship is less apparent in less-exposed work. In the highest-exposure group, employment among people aged 15 to 24 ended below its late-2022 level while employment among those aged 25 to 64 rose. Hours worked by younger people also weakened most clearly in the more exposed groups.
Graduate figures narrow the question further. The employment-to-population ratio for degree holders aged 20 to 24 has fallen more from its late-2022 level than the ratio for people of the same age with lower qualifications. Among higher-education graduates, the youngest group has also fared worse than older graduates. These comparisons describe outcomes for groups, not the job histories of individual graduates or a count of vacancies removed by AI.
A 2025 Australian HR Institute employer survey, cited by Human Resources Director, found 41% of organisations saying AI had increased entry-level roles and 19% saying it had reduced them. Those are employers’ assessments of entry-level positions, whereas the occupation analysis compares employment outcomes and advertisements across jobs; the measures cover different questions and periods. The survey helps explain why exposure does not imply that every junior role is shrinking.
Why exposure cannot establish the cause
The occupation ranking measures how closely work tasks match what current AI can do, not whether a firm adopted AI or changed its staffing because of it. A task can be automated, or the same technology can help an employee perform it faster. That distinction matters for the outlook: replacing part of a job does not necessarily remove the position, while greater productivity could also change demand for other work.
Young workers’ employment was unusually strong in the tight labour market of 2022, giving the later comparison a high starting point. Their outcomes are also more sensitive to ordinary economic cooling than those of older workers. Changes in working-from-home practices may have affected some of the same occupations. These factors leave open how much, if any, of the recent gap is due to AI.
The latest complete occupation-level figures in the staff analysis run to February 2026. More recent data will need to be mapped to a changed occupation classification, and evidence of firms’ actual AI adoption will be needed to connect exposure with decisions about hiring. For people seeking a first foothold in clerical, programming or contact-centre work, fewer new openings can delay entry into an occupation even when most existing jobs remain in place.
Also read:
Related articles


Australia’s Workplace AI Guide Makes Worker Consultation the First Step

Gallup finds U.S. worker engagement below 30% as AI strains clarity

How to verify a remote job before sharing personal information

UK CFOs Expect AI Job Cuts—but Only 6% of Adopters Report Them

AI Hits UK Clerical Roles First—Headcount Cuts Remain at 6%
Subscribe to our newsletter
Get the latest Web3, AI, and crypto news delivered straight to your inbox.