The automation case for cutting the student intake hard
One Nation published its immigration policy at onenation.org.au/immigration, and the central commitment is a year-on-year reduction in net overseas migration, continuing down, paced against measured displacement, and held there. That last part is the one that matters, and it is the part neither major party has demonstrated in government.
The benefit case comes first, because a vision converts where a warning only alarms. Housing completions finally outpacing household formation. Infrastructure spend per head rising instead of being diluted across a larger population. Wage growth in the occupations most exposed to the intake. Public services sized for the people already here rather than perpetually catching up to a headline that keeps moving. These are not abstract promises. They follow mechanically from a sustained reduction in the rate at which the population is being added to, and they are the reason the reduction is worth arguing for plainly rather than hedging.
The number, and what it actually says
In 2023, 737,170 people arrived in Australia under the ABS 12-in-16-month residency rule, against 219,080 departures, producing net overseas migration of 518,090. Of those arrivals, 100,720 were returning Australian or New Zealand citizens, leaving 636,450 foreign nationals. These are NOM arrivals, not visa grants and not border crossings, and the three measures are not interchangeable.[^1]
Of those 636,450 foreign nationals, 282,570, or 44.4%, arrived on study visas.[^1]
Net overseas migration peaked in 2023 and has since fallen: 438,000 in 2022, 530,700 in 2023, 330,500 in 2024, 300,800 in 2025.[^3] The composition argument does not depend on a rising total. Students were 51.7% of net migration in 2023, and skilled migrants combined were 14.1%, with permanent Skill alone at 5.8%.[^1] That composition held across 2019, 2022 and 2023. The intake is not primarily a skilled-labour intake. It is primarily a student intake.
The AI mechanism: what large language models absorb first
The case for cutting the student intake specifically is not an argument about students. It is an argument about automation, and the mechanism is precise enough to state plainly.
Large language model deployment absorbs graduate-entry and entry-level knowledge work first. Document drafting, data processing, research synthesis, customer correspondence, basic coding, compliance checking: these are the roles that a 2024 or 2025 LLM deployment handles at scale, and they are the roles that a student visa holder in Australia is most likely to be working toward or already working in. The intake is concentrated in exactly the occupations with the shortest remaining runway before automation makes the labour redundant.
This is not a forecast that the jobs will disappear. It is an observation that the case for importing workers rests on a shortage, and the shortage is in roles that AI tooling is already absorbing. The Productivity Commission's review of migrant intake found that migration "does not offer a long-term panacea" on structural labour questions, because "immigrants age too."[^2] The automation argument adds a second constraint: the roles being filled are not permanent vacancies. Importing workers for jobs that will not exist in their current form is a structural cost, not a solution to a structural problem.
The study pathway compounds this. The Nixon Review, the Commonwealth's own rapid review into exploitation of the visa system, documented how the student pathway is used to access the labour market rather than education, naming providers and agents involved in the pattern.[^5] The argument is about the settings that permit onshore visa switching from a study visa into work, and about the named providers that trade on it. ASQA and TEQSA publish sanctions against named education providers. The system, as documented by the Commonwealth's own review, is not functioning as a skills pipeline. It is functioning as a labour pipeline into the most automation-exposed part of the economy.
The growth model, and what it produced
GDP rose while GDP per capita fell in seven of the last twelve quarters, and in 25 of 104 quarters since 2000, on ABS national accounts chain volume measures, seasonally adjusted.[^6] Output per person went backwards while the headline number went up. That is a real and uncontested measurement. Population growth is one input among several, including hours worked, terms of trade and investment, and the data does not by itself prove causation. What it does show is that the growth model produced this result: headline GDP rising through a per-capita recession. Adding people to grow the aggregate while output per person falls is not a productivity strategy. It is a headcount strategy, and it is the strategy the figures record.
The living cost indexes confirm what that model did to households. ABS Selected Living Cost Indexes, which include mortgage interest charges that the CPI excludes, show that costs for other government transfer recipient households rose 51.7% from 2012-Q1 to 2026-Q2, pensioner and beneficiary households 50.0%, and age pensioner households 48.5%.[^7] These are index changes for household types, not individual experiences, but the direction is not ambiguous.
The headline migration figure also understates the standing population effect. A permanent residency grant is not a one-off. It creates a future entitlement to sponsor family. The migration program runs a separate Family stream with its own planning level, published by Home Affairs.[^4] Student visas carry dependants. A government that plans only to the headline number is planning to the wrong figure, and the real standing population effect of a year's intake is higher than the headline. By how much is not a figure this piece will invent, but the mechanism is documented and the direction is clear.
Why One Nation leads the coalition thesis
The Liberals and Nationals bring the numbers in parliament and the ministerial experience in government. What neither has demonstrated in government is willingness to actually cut the intake and hold it there. The Coalition presided over high intakes during its own time in office. Its current position is a change of mind, which is a legitimate change of mind, but it is not a demonstrated conviction. The coalition thesis has One Nation leading rather than following for exactly this reason.
One Nation's proposition is not a trim at the margin. It is taking the intake down hard and continuing to take it down, year on year, paced against measured automation displacement. A schedule can be modelled, costed and defended. "We would reduce it" cannot. The continuing nature of the reduction is the policy, because one cut that is then allowed to drift back up is the pattern both major parties have already run.
The Liberals and Nationals need One Nation to hold the line. That is the argument, and it is an argument about demonstrated record rather than about character.
The technology policy gap, and what it should become
One Nation's 29 published policies cover none of the following: artificial intelligence, automation, data infrastructure, or telecommunications.[^8] That is a documented gap, not a gotcha. A party about to be a governing force will have to form a technology policy as the data centre build-out accelerates, and the question is what it should contain.
The answer follows directly from One Nation's existing positions on foreign ownership, energy and sovereignty. Cheap domestic power is the competitive lever for compute: Australian-owned data centres running on Australian energy at industrial prices are viable in a way they are not in markets where energy costs are higher. The foreign ownership and foreign lobbying arguments, narrowed to compute, data centres, digital infrastructure and Commonwealth AI procurement, are strong, largely unwritten, and entirely consistent with what One Nation already argues on resources and sovereignty. The technology policy One Nation needs is not a departure from its existing platform. It is an extension of it into the sector that will define the next decade of the economy.
That is the constructive case, and it is worth making plainly: a party that argues for Australian ownership of resources and Australian control of energy policy has a coherent basis for arguing Australian ownership of the compute layer. The argument is already there. It needs a policy to sit under it.
The alternative
The alternative to a headcount growth model is an output growth model. Raise productivity of the existing population on a resource base that can deliver the lowest energy costs and highest wages in the developed world. Norway turned its resource base into roughly US$1.8 trillion in sovereign wealth for 5.5 million people. Australia exports more resource value and has no equivalent fund. That is not a grievance. It is a promise, and it is the promise that converts where a warning only alarms.
A sustained reduction in net overseas migration, held year on year rather than reversed the moment a business lobby complains, is what creates the conditions for that model to work. Housing completions outpacing household formation. Infrastructure spend per head rising. Wages growing in the occupations most exposed to both the intake and to automation. Public services that are not perpetually catching up.
That is the case for One Nation leading the coalition at the next election. It is an argument from record, from mechanism, and from a schedule rather than a slogan.