OpenAI expands ChatGPT Edu across four Australian universities

OpenAI will extend its ChatGPT Edu platform to more than 80,000 students and staff at UNSW Sydney alone, as part of expanded partnerships with four Australian universities that pair wide AI access with formal rules on when students may use the technology in assessed work.

UNSW Sydney, the University of Sydney, La Trobe University and Adelaide University signed the expanded agreements. The partnerships cover access to ChatGPT Edu, changes to teaching and assessment, and training designed to help students use AI in academic work and prepare for employment.

The commercial stakes are clear from OpenAI's own data. The company commissioned a survey of 1,018 Australian university students and found that 81% use AI several times a week, while 33% use it multiple times a day. Demand for institutional access is high: 79% of respondents wanted more opportunities to use advanced AI tools on real-world projects, and 61% said AI skills would be essential to success in their chosen career.

The survey also documents how far usage has already moved beyond chat. More than a quarter of respondents, 28%, said they had built an AI agent or assistant to help manage their studies, work or personal life. Entrepreneurship interest is strong: 76% said they were interested in starting a company or working for themselves, and 68% said AI would make it easier to start a business.

How large is each university's rollout?

UNSW is extending ChatGPT Edu to more than 80,000 students and staff, with access introduced in stages. The university plans to integrate generative AI foundations and OpenAI's Codex coding tool into programmes and courses, and to give students opportunities to apply AI through hackathons and industry projects.

A UNSW student tied the approach directly to the job market. "It's important to understand what industry is looking for and how AI fits into that," said Nalin Iyer, a Master of IT student at UNSW.

The University of Sydney is also rolling out ChatGPT Edu, alongside training and guidance for students and staff, with access introduced in stages after a trial. The university is additionally establishing an AUD $1 million AI Sustainability Research and Innovation Fund focused on sustainability research.

La Trobe University has moved fastest. All 3,500 of its staff already have ChatGPT Edu access, and the university plans to extend access across its 45,000 students and staff. Its broader strategy includes what it describes as Australia's first Bachelor of AI Engineering (Honours) and Australia's first AI MBA programme.

La Trobe has also created a senior leadership role to oversee the work. "AI should happen with us, not to us," said Phil Laufenberg, Pro Vice-Chancellor, AI, and Chief AI Officer at La Trobe University. "We want students in every discipline to have the skills and confidence to help shape how AI is used in their profession. That means opportunities to build, test ideas and solve real problems, alongside the critical thinking and judgement to use AI responsibly."

Adelaide University is taking a different route. Rather than a broad student rollout, it is focusing on institutional and cyber security uses of AI as it builds systems following a merger. "This is enabling human-centric AI in a thoughtful way, providing us with access to frontier AI models to strengthen our cyber security posture," said Abhishek Singh, Chief Technology and Security Officer at Adelaide University.

What does AI look like inside courses?

Several initiatives embed AI directly into coursework rather than treating it as a separate technology programme. In La Trobe's paramedicine programme, staff use ChatGPT to generate clinical scenarios delivered through the university's learning management system. The scenarios include interpreting an ECG, assessing a possible heart attack and triaging patients after a mass casualty incident. The stated aim is to give students more chances to practise decision-making before they enter professional settings.

"ChatGPT helps us create realistic clinical scenarios so our paramedicine students can practise difficult field decisions. They still need to interpret evidence, explain their reasoning and defend their decisions. It also helps identify course material that needs updating in line with new research, giving us more time for teaching," said Dr Cameron Anderson, Head of the Paramedicine Program at La Trobe University.

In La Trobe's teacher training, a social robot called Fletch simulates conversations with parents. The robot runs on OpenAI large language models hosted through Microsoft Azure and responds with speech, facial expressions and head movements, allowing trainee teachers to rehearse difficult interactions before facing real families.

La Trobe is also making OpenAI's Codex and ChatGPT Work tools available across the institution and is developing an AI Business Studio. The studio is intended to give more than 1,000 students practical experience working as consultants and forward deployed engineers helping small and medium-sized enterprises adopt AI.

Laufenberg said the internal impact was already measurable. "ChatGPT and Codex have helped shorten early systems development and innovation from months to weeks and, importantly, enabled embedded co-design by staff and students," he said.

How are the universities handling assessment?

The expansion of AI access comes paired with more formal rules on when students may and may not use the technology in assessed work. The four institutions are trying to move beyond blanket bans and instead define where AI use fits specific learning objectives.

UNSW has introduced what it calls a Can, Can't, Must framework for assessment. Under the model, educators specify whether AI use is permitted, prohibited or required, depending on the task and the skill being tested.

The University of Sydney is taking a comparable approach. It assesses students' own knowledge in supervised settings while also asking them to demonstrate how they can use AI in open assessments. The model is designed to distinguish unaided individual performance from the ability to use AI appropriately.

La Trobe's support measures include Responsible and Safe AI learning modules for all students, alongside AI skills development embedded in 50 priority first-year subjects. The university is also developing discipline-specific modules for education, law and health, linking AI literacy to professional responsibilities in each field.

Why does this matter beyond Australia?

The four agreements show how a major AI vendor is converting existing student behaviour into institutional contracts. OpenAI's own survey data makes the case: with 81% of Australian university students already using AI several times a week, the question universities face is no longer whether students will use these tools, but under what rules and with what training.

The pattern also signals where the higher-education market is heading. Universities are shifting from prohibition toward structured adoption — funded research calls, degree programmes, leadership roles and assessment frameworks designed around AI rather than against it. La Trobe's compressed development timelines, from months to weeks, suggest the institutions that embed these tools early expect measurable operational returns, not just student satisfaction.

Whether the Can, Can't, Must model and its equivalents can reliably distinguish unaided performance from AI-assisted work will be the test other universities watch as these rollouts mature.