My child has been accused of using AI on schoolwork. What should I do?

The instinct is to pick a side straight away: to defend your child to the school, or to be angry that they cheated. Hold both off for twenty minutes. Ask what the assignment was, how they went about it, and whether AI was involved in any part of it. Ask it as a question you genuinely want the answer to, because a child who expects to be punished either way has no reason to tell you the truth.

Listen for the ordinary middle cases, because they are the most common. A student who asked AI to fix their grammar, or to suggest a structure, or to explain a concept they then wrote about, may honestly not think of that as "using AI". Whether it broke the rule depends on what the rule was, and that is often exactly what is unclear.

AI detectors estimate how likely a piece of text is to have been generated. They do not know, and they cannot. They are known to produce false positives, and they tend to flag writing that is clear, formulaic or plain: a well-drilled five-paragraph essay, a lab report, or the work of a student writing in their second language. OpenAI withdrew its own AI-text classifier because it was not accurate enough, which says a good deal about how far the rest should be trusted.

Most schools know this, and many treat a detector result as a reason to look more closely rather than as a finding. If the concern rests on a score alone, it is entirely reasonable, and not combative, to ask what else it is based on.

Writing leaves a trail. If your child wrote in Google Docs or Microsoft Word online, the version history shows the piece growing sentence by sentence over hours or days, which looks nothing like a block of text pasted in at once. Handwritten notes, an outline, earlier drafts, the sources they read and the browser history from the evening they worked on it all tell the same story.

The strongest evidence of all is your child talking about the work. A student who can explain why they made an argument, what they cut, and what they would change has shown what no detector can measure. Offer this to the teacher directly: they are welcome to ask your child about the piece, in person, without notes.

Ask for a conversation before any decision is final, and go in asking rather than arguing. What was the rule for this assignment? What prompted the concern? What would resolve it? Bring what you have, let your child speak for their own work, and keep the tone on the question you both care about, which is what actually happened.

If the school has a written academic integrity process, ask for it, and ask how AI-related cases are handled under it. If the outcome still rests on a detector score after you have shown the drafts and your child has explained the work, it is fair to ask for the decision to be reviewed.

Then the goal changes, but the approach does not. Help them say so, plainly, to the teacher, and before it is proven. Schools handle a student who comes forward very differently from one who denied it until the evidence arrived, and the lesson your child takes away from that difference will outlast the grade.

Then work out what the pressure was. It is nearly always time, or material they did not understand and did not want to admit to. Both have better answers, and some of them involve AI used well, as a tutor or a sparring partner rather than a ghostwriter.

Whatever the outcome, two habits protect a student from here on. Write in a tool that keeps version history, and keep notes and drafts until the grade is back. And when AI helped with anything, say so in a line at the end of the work: what was used and for what. A student who discloses has nothing to be accused of, and a student with a visible drafting trail has an answer ready if anyone asks.