Academic integrity

Here is an uncomfortable truth about school and university: nobody actually needs your essay. The world already has millions of essays on the causes of World War One. Your essay is not the product. It is the evidence. Assessment exists to prove what you can do, and once you see it that way, the whole AI and cheating question gets much clearer.

Why the struggle was the point

So why make you write the essay at all, if the world does not need it? Because the writing is where the learning happens. When you wrestle a vague idea into clear sentences, you are not recording thinking that already happened somewhere else. You are doing the thinking. The false starts, the paragraph you delete, the moment you notice your argument has a hole in it: that discomfort is the skill forming. Hand the struggle to a machine and the words still appear, but the thinking never happens, in the same way that paying someone to go to the gym for you produces gym visits and no muscle.

Who gets cheated first

If a machine writes your assignment and you submit it as yours, the person marking it loses a little. The classmates who did the work themselves lose a little. But the person who loses most is you, immediately. You paid for the mark with the one thing assessment was supposed to give you: the skill. The certificate now says you can do things you cannot do, and life has a way of testing that claim later, in a job interview, an exam room or a first week of work, when no tool is there to hide behind.

And the loss compounds, because school is a staircase rather than a set of separate rooms. This term's essay assumes the skills built by last term's essay, and next year's subject assumes both. Outsource one assignment and the next one is genuinely harder for you than for the classmates who did the work, which makes the tool more tempting, which makes the assignment after that harder again. Few students ever decide to cheat their way through a subject. They decide once, in a busy week, and then discover the staircase no longer has a step where they left it.

The line between learning and outsourcing

So where exactly is the line? Here is a way to draw it that works in any subject: AI use that leaves you able to do more without the tool is learning, and AI use that leaves you dependent on the tool is outsourcing. The same chatbot sits on both sides of the line depending entirely on what you ask of it.

• Learning: asking it to explain why the quadratic factorises that way, then closing the tab and doing the next five questions yourself.
• Learning: pasting in your finished draft and asking for the harshest critique it can give, then making your own revisions.
• Learning: generating practice questions the night before a test and marking your own answers against its explanations.
• Outsourcing: asking for the essay, the answer or the code and submitting it with your name at the top.
• Outsourcing: feeding it each homework question one at a time, producing a full page of work you could not reproduce tomorrow.

Notice what the outsourcing rows share. In each one the work exists but the ability does not, and that gap is invisible right up until somebody asks you a question about your own work. The case study at the end of this topic shows that exact moment arriving for a student who never saw it coming.

The first rule is know your rules

There is no single rule for AI in education. One teacher bans it outright. Another allows it for brainstorming but not writing. A university unit might require you to use it and document how. The same use of the same tool can be encouraged in one classroom and misconduct in the one next door. So the first move in every subject is boring and essential: read the unit guide or assessment policy, and if it is unclear, ask before you submit. Not knowing the rules has never been a defence, and with AI the rules genuinely differ from room to room.

There is a misconception running the other way that deserves untangling too: that using AI at all is cheating, so the only safe option is pretending the technology does not exist. But cheating was never about the tool. It is about the claim you make when you submit, because your name on the work is a statement that the thinking behind it is yours. A spellchecker does not make that statement false, and neither does asking a chatbot to explain photosynthesis before you write about it in your own words. Submitting the machine's thinking as your thinking does. Keep your eye on the claim rather than the technology and most of the grey areas resolve themselves.

Detection is unreliable. Honesty is not

You may have heard that AI detectors will catch cheats, or that they never work. The evidence says they are unreliable in both directions: they flag honest human writing as machine made, and they miss plenty of AI text, especially after light editing. That cuts two ways. Students have been wrongly accused, which is a real injustice. And students who cheat are gambling that a flaky tool, a suspicious teacher and a follow up conversation about their own work will all go their way, forever. Honesty is the only strategy that does not depend on luck.

Australia takes the supply side of cheating seriously enough to have made it illegal to provide or advertise cheating services to university students, the industry known as contract cheating. And universities have been shifting their defences away from unreliable detectors towards checks that are hard to fake: supervised exams, oral defences where you explain your own work, and short interviews about your methods. The direction is worth noticing. Institutions are not really trying to catch the text. They are trying to test the person, because the person is what the qualification certifies.

What this looks like on a Tuesday night

The real test of all this is not some dramatic moment of temptation. It is 9pm on a Tuesday, you are stuck on question 7, and the chatbot is one tab away. The weak move and the strong move begin identically, because both start with asking for help. The difference is what you ask for. Ask for the answer and you finish faster tonight and arrive weaker at the exam. Ask it to explain the step you are stuck on, then close the tab and finish the question yourself, and you finish a little slower tonight and arrive stronger. Over a term those two habits produce two different students from the same starting point, and neither of them needed to be caught for the difference to be real.

The next lesson covers the other half of honest use: saying so. Knowing the rules and using the tool to learn is most of the job, however the finishing move is a short written note recording what you did, and it is both easier and more protective than most students expect.