Nearly 80% of students use AI. Universities must teach them how.
Students have already made the choice. Nearly 80 per cent of students in Australia use generative AI in some form, and 94 per cent of UK students use it to support assessed work, according to recent surveys. The debate over whether students should use these tools has become obsolete.
Universities now face a different question: will they teach students to use AI well, or leave them to figure it out alone?
From prohibition to literacy
The answer requires moving beyond AI bans and toward structured AI literacy as a core graduate capability. This means teaching students what AI can and cannot do, how to check outputs against evidence, when to use these tools and when not to.
AI literacy is discipline-specific, not a technical add-on. A law student needs different skills than an engineer, who needs different skills than a creative professional. All of them, though, will work alongside AI systems in their careers.
The cognitive trainer versus the forklift
Used poorly, AI becomes a forklift for cognition. Students outsource thinking and mistake fluent outputs for learning. Used well, it acts like a personal trainer for the mind - helping students attempt harder problems, receive faster feedback and build capability through practice.
The difference lies in assessment design. If educators reward only polished text, students will use machines that produce polished text. To encourage deeper thinking, assessments should examine the work behind the output: the questions students asked, the data they used, the assumptions they challenged, the outputs they rejected and the reasoning behind their final answer.
This might mean requiring AI use statements, prompt trails and reflections. It could also mean oral defences, live demonstrations or in-class checkpoints that confirm human learning occurred.
Access matters for equity
Without scaled access to AI tools, universities risk widening existing inequities. Some students arrive having used advanced tools for years. Others have been warned away from them, priced out or unsure what is permitted.
Without a common foundation, institutions create two groups of graduates: those who practised working with AI under expert guidance, and those who experimented privately and anxiously. The same applies to staff. Academic and professional employees cannot model good AI use if left to build capability alone.
Institutions need a coordination layer: setting common frameworks, supporting local AI use and helping experiments become scalable practice.
The market is already moving
Employers signal the urgency. Workers with AI skills command a 56 per cent wage premium on average, according to PwC's 2025 Global AI Jobs Barometer. This is not confined to technology roles - AI is changing every field.
Graduates who stand out will be those taught to use AI to think better, learn deeper and act ethically. Universities have a responsibility to make that capability teachable and visible before students leave.
For educators, this means rethinking how you design learning, what you assess and how you support both students and colleagues. Learn more about AI for Education and explore the AI Learning Path for Teachers to build your own capability.
Your membership also unlocks:
