AI for students: where study ends and cheating begins

Every student is already using these tools. Pretending otherwise helps nobody, so this is the honest version, written for students and for the parents wondering what to think about it.

The line

It is simpler than it is usually made to sound.

Using it to understand something is study. Using it to produce something you hand in is cheating.

Everything else follows from that.

Genuinely good uses

Getting something explained again. This is the best use by a distance. You did not follow it in class, the textbook assumes you did, and asking the teacher a third time feels impossible. A chatbot will explain it fifteen different ways without a hint of impatience.

Ask it to explain like you are twelve. Ask for an analogy. Ask what the point of the concept is. This is exactly what a tutor does, and it is free.

Testing yourself. Ask it to quiz you on a topic, then answer without looking. Being tested is one of the most effective study techniques there is, and almost nobody does it because making the questions is tedious.

Checking your understanding. Explain a concept to it in your own words and ask what you got wrong. Very effective, because explaining something is how you find out whether you actually understand it.

Unstacking a problem. “I do not understand why this code will not run.” “Why is my thesis statement weak?” Ask it to explain the problem rather than fix it.

Planning. Breaking an assignment into steps and working out a realistic order.

Where it becomes cheating

Having it write the assignment. Obviously.

Having it write “a draft” you then reword. This is the one students tell themselves is fine. It is not, and it is generally covered explicitly by school and university policy.

Having it do the specific thing being assessed. If the task is to write code, having it write the code defeats the point. If the task is to analyse a poem, having it produce the analysis does too.

Using it in an exam where it is not permitted.

Two practical reasons beyond the rules

You will be caught more often than you think. Not always by detection software, which is unreliable in both directions. Usually by the far simpler method: work that does not sound like you, that is stronger than your class contributions, or that contains a confident claim about something that does not exist. Teachers who read your writing weekly notice immediately.

The invented sources are what catches people most often. These systems generate references that look completely convincing and do not exist. Marking a bibliography of imaginary books is a very short conversation.

The learning is the point. Sitting with something you cannot do until it makes sense is the actual mechanism by which you get better at anything. Skipping it does not accelerate you, it removes you from the process. The assessment is only a proxy for whether that happened.

That matters most in subjects like Software Engineering, where the HSC includes a written examination and a major project. If AI wrote your project, the exam finds out.

For parents

A few things worth knowing.

Banning it entirely will not work, and it puts your child at a disadvantage against classmates using it well. Every workplace they enter will expect some fluency with it.

The useful conversation is about which uses help. “Did it explain it to you, or did it do it for you?” is a question a fourteen-year-old can answer honestly.

Watch for work appearing too easily. If assignments that used to take three evenings now take twenty minutes and come back with good marks, something has changed and it is worth a conversation rather than an accusation.

Check the school’s policy. Most now have one, and they vary considerably. Some permit AI for research and planning with declaration, some prohibit it entirely for assessed work.

The habit worth building

Whatever they use it for, one rule covers most of the risk:

Never hand in a sentence you could not explain if asked.

If a teacher stops you and asks what that paragraph means, or why you chose that approach, you should be able to answer. If you cannot, it did not come from you, and it should not be on the page.

That test is simple enough for a Year 8 student and rigorous enough for a university thesis.

And one thing genuinely worth learning

Beyond schoolwork, understanding how these tools actually work, where they fail, and how to check them is turning into a real employable skill in itself.

That side of it is covered in the AI skills lessons, and for students interested in a technical career it sits alongside getting into IT properly.