AI Homework Suspicion: What to Do Before Accusing Your Kid

The essay reads too smoothly. The vocabulary isn’t your 13-year-old’s. And it took twenty minutes. If that’s what you’re looking at tonight, you’re not alone: in polling for the UK’s Parent Voice Project, the share of parents who suspect their child is producing AI-generated homework rose from 47% to 57% in under a year. AI homework suspicion is now a common household feeling. But the way you respond matters more than whether you’re right. Harvard researchers have found that suspicion without conversation can push honest kids to hide their best work, and the tools people reach for to confirm a hunch are unreliable. Here is a fair way to find out what happened, and to make it a learning moment either way.

Key Takeaways

• UK polling for the Parent Voice Project found parental suspicion of AI-generated homework rose from 47% to 57% in under a year.
• AI detectors are not reliable enough to settle the question: a 2023 test of 14 tools found none above 80% accuracy, and another found they misflag non-native English writers.
• Harvard’s Center for Digital Thriving found suspicion without conversation leads some teens to deliberately do worse work so they won’t be accused.
• The best evidence of authorship is your child explaining the work: the argument, the choices, the hard parts, and the drafts.
• If AI was used, focus on why, the rules, and what to do next time; most kids who lean on AI are stuck, overloaded or confused about the rules.

Why AI homework suspicion is rising

AI homework suspicion is a parent’s belief, often based on a gut feeling, that schoolwork was produced by a chatbot rather than by their child. It’s increasingly common.

The Parent Voice Project, a UK organization that polls parents through the research consultancy Public First, surveyed about 6,000 parents in 2025 and about 2,000 ten months later. The share who suspect their child is producing AI-generated homework rose from 47% to 57%. Parents in focus groups described confusion: one father said he found out his daughter “was just taking a picture [of her homework] and AI was giving her the answer”; another admitted he’d let his son use it for a subject he was struggling in. The page we read reprints The Times’s coverage of the report; we did not open the original Times article.

US data point the same way from the other side. In Pew Research Center’s 2026 teen survey, 54% of US teens had used chatbots for schoolwork help, and 10% said they do all or most of their schoolwork with chatbots. Pew’s parent survey found many parents don’t know: 51% said their teen uses chatbots, versus 64% of teens.

So the suspicion isn’t irrational. AI use on homework is common. But common doesn’t mean your child did it on this assignment, and how you find out shapes what your child learns.

Why “running it through a detector” isn’t the answer

The instinct is to paste the essay into an AI detector. Don’t treat the result as proof.

AI-text detectors estimate how statistically predictable a passage is, because language models tend to produce smooth, regular text. But plenty of human writing is smooth and regular: formal school essays, careful writers, and kids writing in a second language. The research is consistent:

• Debora Weber-Wulff and colleagues tested 14 detection tools in a 2023 study in the International Journal for Educational Integrity and found them “neither accurate nor reliable,” all below 80% accuracy. They concluded the tools shouldn’t be used as evidence of misconduct.
• A 2023 Stanford study in Patterns by Weixin Liang and colleagues found detectors “consistently misclassify non-native English writing samples as AI-generated.”

California’s model school AI policy puts it in policy form: detectors “should not be used as the sole basis for disciplinary action,” and concerns should rest on “documented evidence of authorship.” If schools shouldn’t accuse on a detector score alone, parents shouldn’t either.

What a false accusation costs

Being wrong is worse than it sounds.

In September 2026, Harvard’s Center for Digital Thriving reported on interviews with 31 teens aged 15 to 19. Teens described fear of false accusation and a climate of “two-way suspicion.” Some said they deliberately underperform so their work won’t look AI-generated, which co-founder Beck Tench called “cognitive withholding.” One student said: “It sucks because I never was going to cheat, but now it feels like I have to.”

That was about teachers, but the dynamic applies at home. A child accused of cheating on work they did themselves learns two things: that good work is suspicious, and that you don’t trust them. Neither helps. We explore this in our article on cognitive withholding.

A fair conversation: what to do before you accuse

This is a sequence, not a script. The goal is to learn what happened, not to win.

Step	What to do	What it tells you
1. Wait until you’re calm	Don’t start the moment you read it	You’ll ask better questions
2. Ask about the work, not about AI	”Walk me through your main argument”	Whether they understand it
3. Ask about choices	”Why did you pick this example? What was hardest?”	Whether they made decisions
4. Look at the process	Drafts, notes, version history, the assignment sheet	How the work developed
5. Then ask directly, without accusing	”Did you use AI for any of this? It’s okay to tell me”	Their account, in their words
6. Check the rules together	What did the teacher allow for this assignment?	Whether there was a violation at all

Step 1: Wait until you’re calm

If you start upset, your child will defend rather than explain. Wait until after dinner or the next morning.

Step 2: Ask about the work itself

Ask your child to explain the essay’s main point in two sentences, or to talk you through how they solved a math problem. A child who wrote it can usually do this, even if imperfectly. A child who pasted it often can’t. This is the same principle teachers use when they follow up with a quick oral check.

Step 3: Ask about the choices

“Why that example?” “Which part took longest?” “What did you change from your first draft?” Real authors remember their decisions and their struggles.

Step 4: Look at the process

If your child writes in a tool with version history, look at it together. An essay that grew over two evenings looks very different from one that appeared in a single paste. Notes, outlines and crossed-out drafts count too.

Step 5: Ask directly, and make honesty safe

Only now ask: “Did you use AI for any of this? I’m not trying to catch you. I want to understand.” Many kids will tell you, especially if they believe the answer won’t end in punishment. Our article on why kids hide their AI use explains how fear drives concealment.

Step 6: Check what was actually allowed

Teachers’ AI rules vary by class and by assignment. Using AI to brainstorm or check grammar may have been fine. Find the assignment instructions or ask the teacher before deciding anything was wrong.

If your child did use AI

Stay curious about why. Stanford’s Denise Pope, whose Challenge Success surveys have tracked student cheating for over 15 years, says students cheat for reasons that have “very little to do with their access to technology”: struggling without help, too much work, assignments that feel pointless, and pressure not to disappoint their families. Each points to a different fix.

Then agree on what happens next. That might be redoing the assignment, telling the teacher together, or setting a home rule like “AI can explain, not write.” If the rules were unclear, help your child ask the teacher. For a concrete conversation guide when a school flags work, see talking to your kid after a watermark flag.

What not to do: accuse first, then look for evidence

Starting with “I know you used ChatGPT” puts your child on the defensive and makes an honest answer harder. If you’re wrong, you’ve damaged trust for nothing. If you’re right, you’ve made the conversation about punishment rather than learning.

What to Watch For Over the Next 3 Months

• Week 4: Make “walk me through it” a normal, low-stakes part of homework check-ins, not only when you’re suspicious. Kids who explain their work routinely get better at it.
• Month 2 red flags: Your child consistently unable to explain finished work, refusing to show drafts, grades on tests falling while homework looks polished, or your child saying they write “worse on purpose” so teachers won’t suspect AI.
• Month 3 self-check: Do you know each class’s AI rules? Does your child disclose AI use without being asked? When you’ve asked about AI, has your child felt safe enough to answer honestly?

Frequently Asked Questions

How can I tell if my child’s homework was written by AI?

You usually can’t tell from the text alone, and detectors are unreliable. The best evidence is your child’s ability to explain the work: the main argument, why they chose examples, what was hard. Drafts and version history help too. Ask calmly, about the work first, before asking about AI.

Should I use an AI detector on my child’s homework?

Not as proof. Independent studies found detectors “neither accurate nor reliable,” and one found they misflag writing by non-native English speakers as AI-generated. California’s model school policy says detectors shouldn’t be the sole basis for discipline. A conversation about the work is far more informative.

What should I do if my child admits using AI?

Find out why: were they stuck, overloaded, confused about the rules, or just taking a shortcut? Check whether the teacher allowed that use. Then agree on next steps, such as redoing it, telling the teacher together, or setting a rule like “AI can explain but not write.” Treat it as a learning moment.

What if I accuse my child and I’m wrong?

Apologize clearly and explain why you were worried. False accusations can teach kids that good work looks suspicious; Harvard researchers found some teens deliberately do worse work to avoid AI suspicion. A sincere apology and a shift to “show me how you work” can repair trust.

Is my child more likely to use AI than I think?

Possibly. Pew found 64% of US teens use chatbots, while only 51% of parents believed their teen did. That doesn’t mean your child used AI on a given assignment. It means talking about AI regularly, before there’s a problem, is worth doing.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. The Parent Voice Project. (2026, August 12). “School homework ‘worse than useless’ due to widespread use of AI” (reprint of The Times coverage of “How Parents Engage with Schools,” polling by Public First). https://theparentvoiceproject.org.uk/2026/08/12/school-homework-worse-than-useless-due-to-widespread-use-of-ai/
2. Tench, B., Weinstein, E., & James, C., with Starks, A., & Konrath, S. (2026). “An AI Policy Isn’t a Playbook.” Center for Digital Thriving, Harvard Graduate School of Education. https://digitalthriving.gse.harvard.edu/an-ai-policy-isnt-a-playbook/
3. Pew Research Center. (2026, February 24). “How Teens Use and View AI.” https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/
4. Weber-Wulff, D., et al. (2023). “Testing of detection tools for AI-generated text.” International Journal for Educational Integrity, 19, 26. https://link.springer.com/article/10.1007/s40979-023-00146-z
5. Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). “GPT detectors are biased against non-native English writers.” Patterns, 4(7). https://doi.org/10.1016/j.patter.2023.100779
6. Stanford Graduate School of Education. (2023). “What do AI chatbots really mean for students and cheating?” https://ed.stanford.edu/news/what-do-ai-chatbots-really-mean-students-and-cheating
7. California Department of Education. (2026). “Model Policy: Artificial Intelligence in Education.” https://www.cde.ca.gov/ci/pl/aipolicy.asp