Your child got flagged for AI. Here's exactly what to do next.

First: do not reply today

The single most damaging thing parents do is answer within the hour, angry, before talking to their child. You lose the ability to change your position later, and if it turns out AI was used, your credibility is spent on the wrong hill.

Acknowledge receipt, say you'll respond within two school days, and stop there.

Second: ask your child, properly

Ask an open question — "walk me through how you wrote this" — not a yes/no one. Then ask them to explain the third paragraph without looking at it. A child who wrote it can. A child who didn't will describe what it says rather than why it's there.

If they did use AI, that changes the entire strategy: you're now negotiating a consequence and a rewrite, not contesting an accusation. That is a far better outcome than being caught defending a false claim.

Third: gather process evidence, not opinions

Detector scores are contested; drafting history is not. Collect, in this order:

• **Version history** from Google Docs or Word — it shows the document being built over time.
• **Outlines, notes and photos** of handwritten planning.
• **Browser or library history** for the sources used.
• **A short recording or written account** of your child explaining the argument in their own words.

Fourth: request a conversation, not an appeal

Most teachers want reassurance, not a conviction. An email that opens with a request to understand the concern gets a very different reply than one that opens with the accuracy statistics of the detector. Keep it to four short paragraphs: acknowledgement, what your child says, the evidence you have, and a request to meet.

Ask one specific question in writing: what does the school's policy say is required to substantiate an academic integrity finding? Most policies state that a detector score alone is not sufficient. Let the policy make your argument for you. The underlying reliability problem is covered in are AI detectors accurate?.

The four phrases that make it worse

• **"The detector is wrong, everyone knows they're unreliable."** True and unpersuasive — it reads as a rehearsed defense.
• **"My child would never do that."** Unverifiable, and it stakes your credibility on something you can't know.
• **"This is harassment / I want this removed from the record."** Escalates a conversation into a case file.
• **"Other kids do it too."** Concedes the point.

Afterwards: make it unrepeatable

Whatever the outcome, install the habit that prevents round two: draft in a document with version history on, keep the outline, and write one line disclosing what AI was used for. That's the disclosure rule at the center of our AI homework rules for families. And if you want a second read on the work itself — in plain language rather than a percentage — paste it into the Homework Check on our homepage.