Why Kids Hide AI Use, and How to Change That

Your kid says the essay is “all mine.” Later you find the chatbot tab still open. Before you decide your child is a liar, consider this: kids hide AI use for the same reasons grown adults do, and researchers have now measured those reasons. People who say they used AI are trusted less than people who don’t say anything. People who let an AI draft their words don’t feel like the authors, yet they still sign their own names. And teens in schools fixated on cheating say they hide AI use, or even do worse on purpose, to avoid suspicion.

None of that makes hiding OK. It does change what works. Lecturing a kid about honesty while the whole system punishes honesty is a losing argument. Changing the rules at your own kitchen table is a winnable one.

Key Takeaways

• Across 13 experiments published in 2025, people who disclosed AI use were trusted less than people who stayed quiet, but being exposed by someone else cost even more trust.
• The “AI Ghostwriter Effect” (Draxler et al., 2024): people don’t feel they own AI-written text, yet most still don’t credit the AI.
• Harvard’s Center for Digital Thriving heard teens describe hiding AI use, and some underperforming on purpose to avoid false accusations.
• Only 27% of US teens say a teacher has ever talked with them about using AI safely (Common Sense Media, August 2026).
• A family norm of “say how, not just whether” lowers the cost of telling and keeps your kid in practice for the disclosure rules schools are adopting.

Why kids hide AI use: three forces, all measured

Hiding AI use means presenting work as fully your own when a chatbot wrote, rewrote or shaped part of it, and not mentioning it. It is not always a calculated lie. Often it is a slide: the kid asks for “a little help,” the help turns into a paragraph, and by the time the paper is due, admitting it feels bigger than it did at the start.

Three separate lines of research explain why that slide happens.

The trust penalty. In 2025, Oliver Schilke and Martin Reimann published 13 experiments in Organizational Behavior and Human Decision Processes. Their finding, in their words: “actors who disclose their AI usage are trusted less than those who do not.” The penalty showed up across tasks, from analysis to artistry, and across roles, from supervisors and professors to creatives. It held whether disclosure was voluntary or mandatory. Kids are not imagining it. Telling really does cost something.

The ghostwriter effect. A 2024 study in ACM Transactions on Computer-Human Interaction by Fiona Draxler and colleagues (two studies, 30 and 96 adult participants) found that “users do not consider themselves the owners and authors of AI-generated text but refrain from publicly declaring AI authorship.” People know, privately, that the words aren’t theirs. They put their name on them anyway.

The suspicion trap. In September 2026, Harvard’s Center for Digital Thriving released An AI Policy Isn’t a Playbook, built on a nationally representative survey of more than 1,000 public school teachers and principals plus interviews with 31 young people ages 15 to 19. Among educators who had faced an AI dilemma, 74% of teachers and 69% of principals described something about cheating. Teens described fear of being falsely accused. Some said they deliberately underperform so their work doesn’t look “too good.” Co-founder Beck Tench calls this “cognitive withholding,” and told Education Week that punitive bans push AI use further out of sight.

Put those together and hiding looks less like a character flaw and more like a reasonable response to bad incentives. That’s the part parents can change.

What the research says about getting caught versus telling

Here is the finding that gives parents their best argument. In the same Schilke and Reimann paper, the trust penalty for disclosing was “comparatively weaker than the effect of third-party exposure.” Translation: if you tell, people trust you somewhat less. If someone else finds out, they trust you a lot less.

That is a concrete, adult-world fact you can hand a 13-year-old. Not “honesty is a virtue,” which they’ve heard, but “the numbers say getting caught is worse than telling, every time they tested it.”

The Draxler study adds a second lever. Participants’ sense of ownership went up when they had more influence on the text. The more of their own thinking went into it, the more it felt like theirs. That suggests the fix isn’t only “admit it.” It’s “put more of yourself in, so there’s less to admit.”

Now the limits, honestly stated:

• Both studies used adults. No one has run these exact experiments on 12-year-olds.
• The trust experiments mostly used vignettes (people rated a described person), not real classroom consequences.
• The Harvard report is qualitative on the teen side. Thirty-one interviews tell you what pressures exist, not how common they are.

The direction is consistent, though, and it matches what large surveys show about how normal AI schoolwork has become. Pew Research Center surveyed 1,458 US teens in fall 2025: 54% have used chatbots for schoolwork help, about one in ten say they do all or most of their schoolwork with chatbot help, and 59% think using AI to cheat happens regularly at their school. When kids believe “everyone does it” and “telling gets you punished,” secrecy becomes the default.

Hiding versus disclosing: what each costs a kid

This table pulls the research into the trade-off your kid is weighing, whether or not they’d put it this way.

	Kid hides AI use	Kid discloses AI use
Short-term trust	Unchanged, if never found out	Some loss (Schilke & Reimann, 2025)
If discovered later	Larger trust loss than disclosing	Nothing left to discover
Sense of ownership	Low, and stays low (Draxler et al., 2024)	Can rise if the kid adds more of their own input
School rules	Increasingly counts as plagiarism (e.g., California model policy)	Meets the rule where disclosure is required
Learning	Adult never sees the real process, so no coaching	Adult can see where the kid needed help
Stress	Ongoing worry about being caught	One awkward moment, then done

The column that matters most is the last row. Hiding isn’t a one-time choice. It’s a running cost.

What schools are starting to require

Disclosure is becoming a formal rule, not just a value. California’s Department of Education published a Model Policy: Artificial Intelligence in Education under Senate Bill 1288. It states that “submitting AI-generated content without clear disclosure constitutes plagiarism and academic dishonesty,” expects students “to disclose their use of AI when submitting work,” and says teachers decide what AI use is allowed assignment by assignment.

Two details matter for parents. First, it’s a model: the CDE says districts can adapt it, and compliance is not mandatory. Your district’s version may differ. Second, the same policy says detection software “should not be used as the sole basis for disciplinary action” and that concerns should rest on teacher judgment and “documented evidence of authorship.” Disclosure protects kids on both ends: it meets the rule, and it creates a record. For more on what happens when a detector flags a child, see our guide on AI cheating accusations and what parents should know before assuming the worst.

The gap is training. Common Sense Media’s August 2026 survey of more than 1,000 teens found 70% use AI for schoolwork, but only 27% say a teacher has ever talked with them about how to use AI safely. And 44% of teen AI users have had a tool blocked at school; of those, 59% switched to a personal device. Blocking moves AI use to the phone in their pocket. It doesn’t make it visible.

How to make telling cheaper than hiding at home

You can’t fix the trust penalty in the wider world. You can remove it inside your house. These tactics are built from the research above.

Make “how did you use it?” a normal dinner question

Ask about AI the way you’d ask about a calculator. “Did you use anything on that? What for?” Asked casually and often, it stops being an interrogation. The goal is a household where mentioning AI use is boring. Boring is safe. Kids hide things that trigger big reactions.

Adopt a disclosure norm: say how, not just whether

“I used AI” tells a teacher nothing. “I wrote the draft, then asked ChatGPT to find places where my argument was weak, and I rewrote two paragraphs” tells them everything. Teach your kid a one-line format: what they did, what the AI did, what they changed. A fuller walk-through by age is in our disclosure guide for citing AI in homework.

Separate the confession from the consequence

If your kid admits AI did too much on an assignment, thank them first, then talk about what to do. If every admission brings a punishment, you are re-creating the exact trust penalty Schilke and Reimann measured. You can still require a redo. Just make it clear the redo is about learning, and the honesty bought them something: your trust and your help.

Push for more of their own input, not zero AI

The Draxler finding is useful here: ownership rose with influence. Instead of banning the tool, raise the bar for what has to come from your kid. Their outline. Their first draft. Their examples. AI can come later, as a critic. A kid who did most of the work doesn’t need to hide the rest.

Model it yourself

The California model policy asks teachers to disclose their own AI use “to model responsible practice and establish reciprocal trust.” Do the same. “I had AI draft that email to your coach, then fixed the tone.” Kids notice when adults hold them to a rule adults ignore.

What not to do

Don’t run their work through an online AI detector and confront them with the score. Detectors produce false positives, the California policy says they can’t be the sole basis for discipline, and a wrong accusation is the fastest way to teach a kid that honesty doesn’t pay. Harvard’s teens described exactly that dynamic. If you’re worried, ask to see drafts and version history, and ask your kid to explain their argument out loud.

What to Watch For Over the Next 3 Months

• Week 4: Is AI coming up in normal conversation without you prompting? If your kid volunteers “I used it to check my math steps,” the norm is taking hold.
• Month 2 red flags: Sudden defensiveness about schoolwork, closed laptops when you walk by, or work that your kid can’t explain in their own words. Any one alone means little. A cluster means the cost of telling still feels too high, so check whether you’ve been reacting harder than you meant to.
• Month 3 self-check: Can your kid write a one-line disclosure for an assignment without help? Do you know your school’s actual AI rule for each class? If both answers are yes, you’ve closed most of the gap the research describes.

Frequently Asked Questions

Is my kid lying if they don’t mention using AI?

Not always on purpose. Draxler’s research shows adults also feel the words aren’t theirs yet still claim authorship, so it’s a common human pattern. But where a school requires disclosure, staying silent can count as plagiarism under policies like California’s model. Treat it as a habit to fix, not proof of bad character.

Will my kid get in trouble for admitting AI use?

It depends on the class rules. Many teachers allow some uses, like brainstorming or proofreading, and ban others. Admitting an allowed use shouldn’t cause trouble. Admitting a banned use may, but research on trust shows being found out later usually costs more than telling first. Help your kid check the rule before the next assignment.

Should I ban AI so there’s nothing to hide?

Bans tend to move use out of sight rather than end it. Common Sense found 59% of teens who hit a school block switched to a personal device. A clear rule about which uses are fine, plus a habit of saying how they used it, gives you more visibility than a ban you can’t enforce.

How do I bring this up without starting a fight?

Start with your own use, then ask about theirs: “I used AI for a work email today. Do you use it for school? For what?” Keep your reaction neutral the first few times. You’re showing that the topic is safe. You can talk about rules once you know what’s actually happening.

What if a teacher already suspects my kid?

Ask what evidence the concern rests on, and gather drafts, notes and document version history. Ask your kid to explain the work out loud. California’s model policy says detector scores shouldn’t be the only basis for discipline. Stay calm with the school; your kid is watching how you handle it.

---

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. Schilke, O., & Reimann, M. (2025). “The transparency dilemma: How AI disclosure erodes trust.” Organizational Behavior and Human Decision Processes, 188. https://doi.org/10.1016/j.obhdp.2025.104405
2. Draxler, F., Werner, A., Lehmann, F., Hoppe, M., Schmidt, A., Buschek, D., & Welsch, R. (2024). “The AI Ghostwriter Effect: When Users do not Perceive Ownership of AI-Generated Text but Self-Declare as Authors.” ACM Transactions on Computer-Human Interaction. https://doi.org/10.1145/3637875
3. 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/
4. Education Week. (2026, September 30). “How AI Suspicions Are Undermining Student-Teacher Relationships.” https://www.edweek.org/technology/how-ai-suspicions-are-undermining-student-teacher-relationships/2026/09
5. 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/
6. Common Sense Media. (2026, August 18). “Teens in the AI Era: Schoolwork and Skills That Matter.” https://www.commonsensemedia.org/research/teens-in-the-ai-era-schoolwork-and-skills-that-matter
7. California Department of Education. (2026). “Model Policy: Artificial Intelligence in Education.” https://www.cde.ca.gov/ci/pl/aipolicy.asp