The federal Department of Employment and Workplace Relations released a report on 8 July that reached a blunt conclusion: artificial intelligence is not, so far, driving broad job losses across the Australian labour market.
The report, titled AI and employment in Australia, tracks how occupations have shifted since generative AI tools went mainstream in late 2022. Its headline finding is that the feared collapse in entry-level and youth work has not shown up in the data.
Instead, jobs for people aged 20 to 24 have grown slightly faster than for those aged 25 and over since the launch of ChatGPT. Software developer roles, often cast as prime targets for automation, are up 25 per cent since November 2022.
As Startup Daily reported, the picture that emerges is one of resilience rather than rupture, at least for now.
What the numbers actually say
The report leans on a straightforward comparison. It splits occupations into those most exposed to AI and those least exposed, then measures how each group has grown.
Employment in the most AI-exposed occupations grew 5.6 per cent since November 2022. The least-exposed roles grew 9.5 per cent over the same period.
That gap matters. It is not a fall, and no category is shrinking outright, but AI-heavy work is expanding more slowly than the rest of the market.
Clerical and administrative jobs, long flagged as the most automatable, sit on the slower side of that line. The broader labour market remains strong by historical standards, with unemployment at 4.4 per cent as of May 2026 and participation near record highs.
In a statement carried by The National Tribune, Employment and Workplace Relations Minister Amanda Rishworth framed the findings as cautiously positive. “AI is changing how we work, but it also has real potential to lift productivity, build skills and create new opportunities,” she said.
The minister also committed to supporting workers “with the skills, training and pathways needed to adapt and benefit” as adoption spreads.
A quiet reshuffle, not a crash
The department itself does not call this an all-clear. Coverage by HRD Australia notes the report found no common employment trend across the jobs most exposed to automation, and that DEWR intends to keep monitoring whether these patterns hold as AI capability improves.
That caution echoes earlier warnings from the Productivity Commission, which last year flagged a “painful transition” for workers whose roles are made redundant, even as it projected sizeable long-run productivity gains from AI adoption.
The commission singled out administrative and clerical roles as the most exposed to displacement, while IT and digital services stood to gain the most. The DEWR data is consistent with that split, just earlier and milder than some feared.
Our read: the more useful signal here is not the absence of job losses but the shape of the growth. When AI-exposed roles add jobs at little more than half the rate of everything else, that is a market quietly rerouting itself rather than one in crisis. The pain, if it comes, is likely to look like fewer new openings and slower hiring in specific functions, not mass redundancies that make headlines. That kind of slow squeeze is harder to see and harder to legislate against.
Why it matters
For Australian workers, the report is a rare piece of hard local evidence in a debate usually dominated by overseas layoffs and vendor hype. The youth and graduate numbers are especially significant, because entry-level roles were the ones economists most expected AI to erode first.
For policymakers in Canberra, the finding buys time but sets a test. The government has tied its response to skills and training, which only works if that investment lands in the occupations where growth is stalling, particularly clerical, administrative and other routine cognitive work.
For local industry, the 25 per cent jump in software developer roles cuts against the assumption that coding assistants would thin the ranks of engineers. It suggests AI is, at least for now, expanding demand for people who can build and deploy it rather than replacing them.
There is also a sovereignty dimension. A workforce that grows its AI-building capacity, rather than just its AI-consuming capacity, is better placed to keep more of the value onshore as the technology matures.
The caveat sits in the report’s own framing. This is a snapshot of the first few years of mainstream generative AI, taken before the most capable agentic systems are widely deployed in Australian workplaces. Whether the resilience holds is the question the next report will have to answer.
Sources: Startup Daily, The National Tribune (Ministers’ Media Centre statement), HRD Australia, HRD Australia on the Productivity Commission, Department of Employment and Workplace Relations.
