International education in Australia will be argued over intensely between now and the next Federal Election.
Even though many universities rely on international students to support their operations, it appears that in many instances, a large number of these students face severe linguistic, and by extension academic, challenges to make the most of their studies here.
In order to deal with this, academics from six Australian universities have come together to tackle this issue head-on – an initiative of staff working across university boundaries, brought together by our combined determination to improve educational experiences and outcomes for international students. This is not an initiative of our universities, but of staff teaching international students every day and seeing a need to do better.
The evidence we have collected about supporting international students is far from reassuring. In a survey of some 430 international postgraduates at a university in South Australia, 77% reported difficulty reading academic research articles and 71% found writing academic essays challenging. Support was available and widely subscribed, with 77% of these students using the central learning support service, and 41% using Studiosity; yet 75% still felt their academic communication skills needed improvement (Botha et al., International Journal of Applied Linguistics, 2026).
These findings suggest that access to support alone may not be enough: we need to better understand what kinds of support work are required, for whom, and in which disciplinary contexts. We need to examine how to connect students to support, and also why students are not getting what they need from existing services. Providing a service and then blaming the student for not accessing it is not a path to positive learning outcomes.
We have also seen that where English and academic literacy instruction is embedded in disciplines, outcomes shift and students transition better. A recent pilot study showed that a compulsory postgraduate STEM subject focused on discipline-specific literacy and team-based learning, lifted confidence in international students’ oral presentations from 73% to 90%, with the cohort’s average GPA rising from 4.13 in 2023 to 4.45 in 2025 (Tan & Botha, Educational Studies, 2026). This is encouraging, but we desperately need to expand the evidence base across other institutions and contexts.
Generative AI (GenAI) has recently upped the stakes in this conversation. In a recent study, and across 462 students and four years of academic integrity case records, 50% of international students at a university reported using GenAI at least half the time in their studies, compared with 19% of domestic students using it at least half the time.
The difference widens where tasks become more language-dependent, such as translating between languages, explaining a concept in simpler English and improving grammar or academic tone.
Over the same period the number of reported integrity cases at the institution has risen dramatically, with international undergraduates referred at roughly twice the domestic rate. However, 62.4% of the 2025 cases were found to close without a misconduct finding (Botha, Tan & Appukuttan, Educational Studies, 2026).
Currently, policy largely governs AI as an integrity risk, but many students use it as a language/academic literacy scaffold, and international students often carry the cost of using it as a language learning support mechanism. Policies cannot distinguish learning scaffold from misconduct. For many multilingual students, GenAI is therefore emerging not simply as a shortcut or “coping mechanism”, but as a form of language and academic-literacy mediation, raising important questions about both its affordances and its risks.
Our program brings together staff from Flinders University, the University of Adelaide, Curtin University, the University of Queensland, Macquarie University and Monash University, across five states, organised around STEM and non-STEM clusters, with deliberately varied governance and support models.
It is a nationwide program to explain international students’ academic literacy in detail and link this to disciplinary knowledge-making.
It is also the first study in Australia to test embedded discipline-specific pedagogy causally and at scale, linking what students report to what institutions record in grades, completion and attrition through a multi-site stepped-wedge trial. We intend to produce a rigorous and nationwide measurement of the effect of GenAI on language and academic literacy, integrity and equity, within a framework that treats belonging and equity as measured outcomes.
The stakes are not only pedagogical. International education is considered among Australia’s largest exports, and its social value rests on retention, completion and the quality of our graduates. On the one hand, better completion and lower attrition protect export revenue and the skilled-migration pipeline. On the other, we argue, an anti-deficit approach reduces linguistic inequity and builds belonging; and an open national dataset, a pedagogy framework and GenAI guidance provide every university and agency evidence-based tools for teaching quality, assessment integrity and responsible use of AI.
The sector does not currently lack opinion about international students. But we do lack evidence. It’s time to build an evidence-based understanding about what works for these students and how best to support their transition to, and through, Australian higher education.