Australia's ABS finds AI barely moved job vacancies in August

Australia's employers report little effect so far

The numbers come from new questions the Australian Bureau of Statistics (ABS) added to its quarterly Job Vacancies Survey, a response to public and policy interest in AI.

Sean Crick, ABS head of labour statistics, said the data showed AI has had a limited impact on vacancies to date. Four per cent of businesses reported that AI had reduced the number of vacancies in August, while less than one per cent reported an increase. Looking ahead, about two-thirds of businesses said they did not expect AI to affect their vacancies over the next 12 months, 8% expected a decrease and 1% expected an increase.

For jobseekers and regulators, the survey says the AI shakeout has not reached the vacancy counts yet. It records what businesses say about their own hiring, not what employers post online. The ABS questions cover August and the next 12 months.

Total job vacancies across Australia stood at 325,000 in August, down 0.9% from May 2026. Private sector vacancies fell 2% to 286,300, while public sector vacancies rose 7.8% to 38,700.

Sector splits cut both ways

Information media and telecommunications had the highest proportion of businesses reporting a decrease in vacancies due to AI, at 16%. Wholesale trade and manufacturing followed. The two published accounts of the release differ on the exact figures: HCamag puts wholesale trade at 8.1% and manufacturing at 6.1%, while the ABS release as carried by Mirage News says eight per cent and six per cent.

The accounts also disagree on where AI added vacancies. Mirage News reports that 10% of businesses in electricity, gas, water and waste services saw an increase. HCamag reports a 16% increase for that sector.

Public administration and safety recorded 4% of businesses reporting an AI-driven increase, and health care and social assistance 2%. By location, the Australian Capital Territory logged the biggest quarterly drop in vacancies at -21.9%, followed by the Northern Territory at -8.0%.

The IMF measures a 10% fall in demand

The International Monetary Fund's (IMF) latest Regional Economic Outlook for Asia and Pacific drew on 40 million online job postings across 19 Asian economies. It found employer demand in occupations more exposed to AI has fallen by nearly 10% since ChatGPT launched in November 2022.

The decline has stabilised despite rapid improvement in AI capabilities, but the report says the effects of reduced hiring are likely to unfold only over time. US data points in different directions. The St. Louis Federal Reserve found a 0.47 correlation between AI exposure and unemployment increases across occupations from 2022 to 2025, with computer and mathematical roles carrying roughly 80% AI exposure. Yale's Budget Lab found a slight employment increase in high-exposure roles through late 2025. Surveys cited in the same report found 37% of companies planning to replace workers with AI by the end of 2026.

The IMF cautions against reading the stabilisation as evidence that displacement risks have receded.

IIT Madras says the unit of change is the task

A white paper from the Centre for Responsible AI at IIT Madras argues the binary question is the wrong one.

"AI rarely replaces a job wholesale; more often it automates a bundle of tasks within a job, and the size of that bundle varies with task structure, sector, regulatory setting, and, distinctively in India's case, the economics of AI access itself."

The paper proposes a four-level framework of task, job, occupation and labour-market outcome, and splits Indian job impact into two regimes. In the first, AI has automated most of the skill bundle and the human is reduced largely to validation. In the second, the human remains the primary driver and AI works as a productivity tool.

It draws on Indian evidence from business process outsourcing, manufacturing, MSMEs, sanitation, financial services, law and software engineering. Five recurring reasons for automation emerge: cutting costs, reaching frontier-grade capability, meeting regulatory requirements, improving safety, and fixing service quality after a failed automation attempt.

Its three-pronged recommendation is to keep encouraging AI diffusion, invest in India-specific labour-market research, and pair adoption with structured worker-transition pathways. The authors also warn that as AI moves from individual tasks towards workflow-level execution, the unit of change may shift from the task to the workflow.