Fifteen universities move to make AI disruption education-led
Fifteen Australian universities are beginning work this week on a new AI coalition designed to move the sector from discussion to action.
The newly established HEDx AI Innovation Coalition brings universities together around one practical test: each institution will take on a substantive AI project and work it through to a live result by November. AI is already disrupting higher education; the question is whether the next phase will be led by educators, or shaped around them by vendors, platforms and students working outside institutional systems.
No university can answer that question alone.
The pace, scale and complexity of AI means the sector needs to build together: sharing knowledge, testing live projects and learning in public before the future of higher education is defined on its behalf.
HEDx CEO and founder Martin Betts has framed the challenge directly: the future of AI in education must be built by universities, not simply around them.
The coalition starts with memory
The coalition does not begin with a blank page. Last week, HEDx launched a new AI agent built by UNE’s LabNext70 on five years of higher education sector knowledge.
HEDx has become one of Australia’s richest archives of higher education thinking, built through podcasts, transcripts, papers, newsletters, book chapters and conference sessions. But that knowledge has been spread across platforms, formats and years of output.
The HEDx AI Agent gives the network a way to interrogate its own collective intelligence without needing to read, watch or listen to everything HEDx has ever produced. For universities beginning live AI projects, that matters. Before deciding what to build, they can ask what the sector has already learned. In this coalition, the HEDx AI Agent becomes the first proof point: this is already moving from talk to infrastructure.
Action, not another report
The coalition draws on Educate Ventures Research’s Shape the Future coalition work in the UK, where almost 1,000 schools and colleges have been actively exploring AI use.
Professor Rose Luckin CBE, Founder and CEO of Educate Ventures Research and Professor Emerita at UCL, said: “We have seen this work with almost a thousand schools and colleges in the UK. Adopting AI tools is only the beginning. The harder question every university faces is what we teach, how we assess it, and whether students keep thinking for themselves when the machine offers to do it for them.”
HEDx and EVR are now bringing that coalition model into Australian higher education. HEDx is convening the work. EVR brings the evidence and evaluation lens. Participating universities bring live institutional problems. UNE’s LabNext70 is contributing practical infrastructure. The point is to avoid the usual pattern: sector anxiety, isolated experiments, duplicated effort, then another report.
This coalition starts with projects. Institutions will build, share progress with the cohort, and feed findings into a shared evidence base the sector can learn from. CEO and Founder of HEDx Martin Betts, said: “This is a resource that no single institution could commission alone. The point of the coalition is to share the sector’s bandwidth of innovation, so universities can build and learn faster together than any one of them could in isolation.”
Madgwick Studio gives universities somewhere to build
The coalition’s second piece of infrastructure is Madgwick Studio. Developed by UNE’s LabNext70, Madgwick Studio gives universities a governed environment to design, test and trial AI-supported learning models. For coalition participants, it means they can start building immediately.
Its deeper proposition is about more than technology. Mass higher education has scaled through large cohorts, stretched academic time and increasingly anonymous learning environments. The student-academic relationship has often been the first casualty of that scale.
Madgwick Studio is built around a different idea: AI handles the scalable work - content delivery, first-pass feedback, adaptive scaffolding and guided practice - so academics can return to the human work of thinking with students, challenging them, mentoring them and knowing them.
AI handles the scale. Humans handle the meaning. UNE Chief AI Officer and Director of LabNext70 Associate Professor Aaron Driver, said: “Higher education did not lose its value because academics stopped mattering. It lost its way when scale made the academic relationship harder to access. AI gives us a chance to fix that. Madgwick Studio is designed so AI handles the scale and humans handle the meaning.”
Equity is the test
The larger test is whether AI can help universities widen access to expertise, feedback and support. That question is landing in a live policy moment. From 1 January 2026, needs-based funding became part of Australia’s higher education funding system, and further Universities Accord legislation was introduced in late June to expand access for students from disadvantaged and regional backgrounds.
The equity gap remains clear. In 2024, regional and remote students made up 19.2 per cent of enrolled domestic students, compared with 27.1 per cent of the Australian population. AI could widen that gap if its benefits flow first to students and institutions already best placed to use it. Or universities could build deliberately and use AI to make academic support more available to the students who need it most. That is the real test of the coalition.
Already tested in public
The HEDx and LabNext70 partnership was tested at the HEDx Disruption Through Connections conference at UTS Sydney in June. LabNext70 presented Madgwick Studio live and ran a hands-on workshop where participants built inside the platform. The HEDx AI Agent was also tested across the day, handling thousands of searches, while the infrastructure recorded zero downtime and zero server errors. The demonstration showed that universities can start testing the future of learning now, with real users and working systems.