How to Reference AI in Schoolwork

Quick answer

Referencing AI in schoolwork means disclosing three things plainly: which tool you used, what you used it for, and how much it shaped the final work. If your school has a required format - a cover sheet, a set statement, a particular citation approach - use that first; schools are increasingly specifying one, and following it beats guessing. Where no format exists, a short disclosure statement at the end of the piece is the safe default: name the tool, describe the task you gave it, and say what you checked or changed afterward. This isn't paperwork for its own sake. It's the transparency principle the Australian Framework for Generative AI in Schools is built on, and it's what turns AI help from a hidden risk into an honest, assessable choice.

Why disclosure matters more than the tool itself

The question most families get stuck on - "was I allowed to use AI at all?" - is usually the wrong one. Most schools are not asking whether a student touched a chatbot; they're asking whether the student was honest about it. Undisclosed AI help, on a task where it wasn't clearly permitted, is what turns ordinary use into a problem. Disclosed help, even substantial help, is usually just help.

This is the same logic behind the Australian Framework for Generative AI in Schools, which lists transparency among its six guiding principles. The framework doesn't treat AI use itself as the risk to manage; it treats hidden AI use as the risk, because hidden use is what a teacher can't fairly assess. A disclosure statement is the practical, everyday version of that principle: it turns an invisible choice into a visible, gradeable one.

What to actually state: the three-part disclosure

A good disclosure answers three questions, in plain language, without padding.

1. Which tool did you use? Name it - ChatGPT, Claude, or whichever assistant it was. Vague phrasing like "I used AI" invites the suspicion that specific phrasing avoids.
2. What did you use it for? Be precise: brainstorming, checking grammar, generating practice questions, drafting an outline, writing full paragraphs. The task matters more than the tool.
3. How much did it shape the final work? Did you rewrite what it gave you, check it against a source, use it only as a starting point? This is the line between "AI helped me think" and "AI did the thinking".

Answering all three, briefly, is usually enough. A disclosure statement doesn't need to be an essay of its own - two or three honest sentences do the job.

Follow the school's format first

Where a school has specified a citation or disclosure format - a cover sheet to sign, a required sentence, a particular style guide's approach to citing generative AI - use it exactly, even if it feels clunkier than writing your own. A format the school set is not optional guidance; it's the rule for that assessment, and following it removes any ambiguity about whether disclosure happened.

If no format exists, the safe default is a short statement placed at the end of the piece, before references, under a heading like "AI use" or "Acknowledgement of AI assistance". When in doubt, ask the teacher directly before submitting rather than guessing - a two-line email answered in advance is worth more than a paragraph of justification after the fact.

Sample disclosure statements

The wording should always be specific to what actually happened, but these show the shape a disclosure takes across common task types.

Notice what each of these does: names the tool, states the exact task, and clarifies what the student did afterward. That last part - what happened after the AI response - is usually the part that actually matters to a teacher assessing the work.

What happens if you don't disclose

Undisclosed AI use on a task where it mattered doesn't just risk a mark. It changes how a teacher reads everything else a student submits, because trust, once it's a live question, colours the whole relationship. It's also unnecessary risk for very little upside: disclosed, well-used AI help is rarely penalised on its own, while concealment discovered later - through a mismatched writing style, an inconsistent process, or an inability to explain the work - tends to be treated far more seriously than the original use would have been.

Common mistakes with AI disclosure

• Being vague. "I used AI to help" tells a teacher nothing and reads as evasive even when it isn't.
• Disclosing only after being asked. Voluntary disclosure lands very differently to disclosure prompted by suspicion.
• Assuming no policy means no rule. If a school hasn't specified a format, ask - don't assume silence means anything goes.
• Copying a disclosure statement that doesn't match what happened. The statement has to describe the actual use, not a generic template.
• Treating disclosure as a loophole. Naming AI use doesn't excuse submitting work you can't explain or defend - see how reliable AI detection tools actually are for what happens when that gap gets noticed.

The recommendation: treat disclosure as the default, not the exception. State the tool, the task, and what you did with what it gave you, follow your school's format if one exists, and ask before you guess. It takes three sentences, and it is the single habit most likely to keep an honest student out of an integrity conversation they didn't need to have.