New Reserve Bank of Australia research finds no clear evidence that AI has yet caused significant economy-wide job losses or materially lifted productivity, but it flags tentative signs of weaker outcomes for younger workers in occupations more exposed to AI.

To look for emerging effects, the analysis compares labour-market outcomes across occupations according to their potential exposure to AI tasks, using November 2022—when ChatGPT became publicly available—as a benchmark.

Economist Mike Major’s analysis uses an exposure measure from Jobs and Skills Australia; software developers, clerical workers and administrative staff are among the occupations most exposed.

Since November 2022, new job advertisements have slowed somewhat more in highly exposed occupations, with general clerks, software and applications programmers, ICT business and systems analysts, call-centre workers and receptionists accounting for much of the decline.

The research also finds employment growth has been stronger in less-exposed areas, particularly for younger workers, suggesting possible early adjustment in the labour market.

Sarah Hunter said roughly three-quarters of firms reported no AI-related change in headcount over the past year; about 12 per cent were hiring more slowly or leaving roles unfilled, 8 per cent had cut staff and 6 per cent had added staff.

The findings are not proof that AI caused these patterns: exposure measures indicate which tasks could be affected, not whether firms have adopted AI or jobs will disappear, and broader labour-market changes can also shape outcomes.

The analysis is descriptive, with the latest complete occupation data covering February 2026, and the Reserve Bank of Australia says it will refine its work as new labour-market information becomes available.

AI-related terms have surged in job advertisements, but that does not establish that AI is replacing other labour; some employers may simply be using fashionable terminology without changing the skills they seek.

AI can automate some tasks, support workers in others and create new roles; estimates cited by Hunter suggest about 4 per cent of Australian workers are highly exposed to automation, compared with 31 per cent whose work is highly exposed to augmentation.

AI is already helping workers complete some tasks faster, but broad productivity gains have yet to materialize; firms may need to invest and reorganize before benefits spread through the economy.

Hunter said AI investment may add to inflation in the near term by boosting demand while supply is constrained, particularly through data-centre construction and related spending, but expanded capacity and productivity gains could ease bottlenecks and price pressures over time.