Students just need to know where they stand on AI

When I went back to university last year after a sabbatical year it felt like walking into a new world.

In the space of a year, AI had gone from something a few friends used casually to something I’d see on almost every screen in the library. I quickly felt like the odd one out for not using it – and I wasn’t sure whether I was even allowed to.

I’ve since realised that feeling isn’t unusual. As an education officer at my students’ union, I hear versions of this all the time. Students are using AI in lots of different ways, and most of them are trying to do the right thing but what they’re missing isn’t rules, it’s confidence.

There has been a genuine move in the sector away from trying to catch students out and towards being clearer about where AI is and isn’t permitted. But the experience on the ground still depends heavily on which course, module, or lecturer you happen to have.

Some lecturers actively encourage students to use AI to make sense of assignment briefs and marking rubrics. Others set oral assessments, or ask you to show how you completed your work, precisely so you can’t lean on it.

The variation in itself isn’t the problem – different disciplines need different approaches but the problem is when nobody explains the variation. If one module says yes and the next says no, and nobody explains why, it starts to feel arbitrary rather than principled, and students end up second guessing themselves.

Pause and panic

When I’m chatting to students, the moment I hear most about is the one just before you hit submit on an assignment.

Students use AI at various points in their learning – organising notes, clarifying a brief, checking their understanding – and by the time they reach the end, many genuinely don’t know whether what they’ve produced will be seen as their own. When I was a student a few months ago, I lost count of the friends who asked me some version of “am I going to get done for this?”

And that fear and assumed guilt starts to build. AI can bypass the learning process – it takes you from A to B to a polished final output. So students are left asking not only “is this allowed” but “does this work actually show what I know?” And these same students are frustrated that they’re not confident they can demonstrate learning.

Students who have used AI entirely legitimately feel this just as much as anyone, and fear of being wrongly accused does nothing to encourage honest, open use. What students need is an official checkpoint – a chance to check their work, reflect on how it was produced, and feel confident they can stand behind it. That is a far better foundation for academic integrity than suspicion after the fact.

For my dissertation, I chose not to use AI. I loved my topic, I had a brilliant lecturer, and – crucially – I had the time. But I know that not all of my peers did. For many students the mark on their transcript is what matters most, and people will take the most efficient route to it.

Which is why we have to ask bigger questions: what is assessment actually for? Is the “thing” a student hands in an output for a grade, or evidence that they have learned something?

In my final year I had varied assessment – an oral assessment, which I loved because I knew my topic inside out, and group work, where you build the human skills employers keep telling us matter. The new formats don’t make AI disappear, students will still use AI, but the ways they use it to support learning rather than bypass it is the difference.

Changing the format without building students’ confidence, and without explaining why, just moves the anxiety somewhere else.

The race to keep up

There is a lot of good in how some students use AI tools.

Conversations gathered by our students’ union found that students, particularly neurodivergent learners, commuter students, and international students, are using it as a study support tool, to ask the questions they might feel too nervous to ask, or when they can’t easily access academic help. But we know that isn’t a long term fix. AI is just filling a gap because support can’t always be there, for example at eleven o’clock at night when some students might work best.

Some students arrive at university having been taught to use AI at sixth form or college or have used it on their placement, and some have no experience of using the tools at all. International students are navigating a new culture and a hidden curriculum on top of a technology nobody has fully got their head around. Students with more time and money can keep up with new models and paid tools while students juggling work and cost of living pressures can’t. If access to better quality AI tools becomes more expensive, those students risk being left behind.

Nobody has all the answers – not students, not staff, not employers and that’s nobody’s fault.

But it means this has to be worked out collectively, with students in the room from the start rather than consulted once decisions have already been made. Clear guidance, explained reasoning, and support that helps students feel confident in their own work would go a long way.

Students don’t need catching out but they do need to know where they stand.