In a piece I co-authored with Meena Jha (CQU) for Times Higher Education Campus, we argued that generative AI has turned teachers into auditors. It is a provocative claim but one grounded in what many educators are living right now. This post takes that argument and brings it home. What does it actually look like at UTS, and what can we do about it?
Here is a question worth sitting with: when did marking start feeling like detective work?
If you have found yourself spending more time scrutinising how a student produced an assignment than what they actually learned, you are not alone. Across universities worldwide, and right here at UTS, Generative AI has significantly changed how students produce their work. It has also changed the very definition of what educators are expected to do, adding new responsibilities that were never part of the job description. For example, interpreting AI policies on the fly, judging whether a student’s reflection is genuine or performative, and making difficult calls in grey areas that no policy quite covers.
From guide to gatekeeper
Think about why you became a teacher. It was not for the marking. It was not for the policy documents. It was for the student who emails you six months later to say something you taught them changed how they see the world. It was for the tutorial that runs ten minutes over because the discussion got too good to stop. It was for that moment when you can see, right in front of you, that something has clicked. That is what teaching is. That is what makes it worth it. That joy is under pressure.
When a student submits an assignment today, many educators are no longer asking only, did they learn this? They are asking: did they write this? Did they think this? Is this reflection genuine? Was the AI a scaffold or a ghostwriter? These are not simple questions and answering them across dozens or hundreds of submissions is exhausting.
In my own postgraduate teaching, I required students to submit GenAI usage templates alongside weekly tasks, including tool declarations, transcripts and reflective commentary. On paper, this creates transparency. In practice, it transfers a significant interpretive burden onto the educator. The marking does not get lighter. It gets different, and often harder.
The gap between policy and practice
UTS has worked hard to develop a clear, principled approach to GenAI. Our position of effective, ethical engagement is the right one. We are not banning these tools, and we are not pretending they do not exist. UTS has developed guidelines, Canvas Commons templates, academic integrity resources, and a Library Generative AI Study Guide that can be embedded directly into subject sites.
But even the best policy hits a wall when it meets the reality of a Wednesday afternoon tutorial with 40 students and three different interpretations of what “permitted use” means.
Policies are, by necessity, broad. Classrooms are messy. The gap between what the policy says and what it means for a specific task, in a specific discipline, with a specific cohort, is currently being filled by individual educators, largely alone, in real time. That is not sustainable, and it is not fair.
What is actually at stake
When the primary lens through which educators view student work becomes one of suspicion rather than curiosity, trust erodes on both sides. Students feel monitored, educators feel like compliance officers, and the relationship that makes great teaching possible starts to break down.
Trust is not a bonus feature of good teaching. It is the whole point. When students feel safe, they take risks. They ask the questions they are embarrassed to ask. They engage with feedback instead of just accepting a grade and moving on. But when they feel constantly watched and second-guessed, they stop doing all of that. Instead, they do just enough to get through. They perform compliance rather than demonstrate understanding. And that is the opposite of everything we are trying to achieve.
So what do we do about it?
The answer is not to ban AI or pretend it does not exist. It is to teach differently. Here are four shifts that are already working at UTS and beyond.
- Bring AI into the room. Run activities that use it. Discuss its outputs together. Ask students to critique what it produces. When GenAI becomes part of the conversation rather than something done in secret, it stops being a threat and starts being a teaching tool.
- Design tasks that require students to think, not just produce. Oral components, in-class activities and iterative drafts are much harder to outsource and far more revealing of what a student actually understands. The goal is not to catch students out. It is to make their thinking visible.
- You do not have to figure this out alone. Talk to your peers, share what is working and what is not. At UTS, learning designers and faculty librarians are also here to help. The UTS Library Generative AI Study Guide, Education Express resources, and the Canvas Commons templates for academic integrity and GenAI are a good place to start, and both can be embedded directly into your Canvas site.
- Talk to your students about it openly. Not as a warning but as a real conversation about their learning, their future careers, and what these tools can and cannot do for them. Students who understand why a boundary exists are far more likely to respect it.
The bigger picture
UTS is a university built on the idea that technology and humanity go together, that innovation and ethics are not in opposition. That is exactly the mindset needed right now.
Generative AI is not going away. The students in our classrooms will spend their careers working alongside these tools. What we teach them about how to engage with AI, critically, ethically and with genuine intellectual ownership, may matter more than almost anything else we cover.
That is not a burden. It is an opportunity. But only if we step back into the role of educator, not auditor.
This post draws on research published in Australasian Journal of Educational Technology (AJET).
Great piece Amara thanks