Only 27% Were Taught: How to Teach Teens Safe AI Use at Home

Seven in ten teens use AI for schoolwork. Twenty-seven percent say a teacher has ever discussed how to do it safely. Those two numbers come from the same survey, published by Common Sense Media on August 18, 2026, and the gap between them is the whole story. If you want to teach teens safe AI use, the school probably isn’t doing it, and the tools your kid is using assume somebody already did. This is a six-week plan to close that gap at your kitchen table, twenty minutes a week, without becoming the AI police.

Key Takeaways

• Common Sense Media’s August 2026 survey of 1,017 teens found 70% use AI for schoolwork, 63% of those get answers from it, and only 27% say a teacher ever discussed safe use.
• “Safe AI use” is five learnable habits: verifying output, protecting personal data, disclosing use, keeping emotional distance from the tool, and knowing when the tool is doing the learning for you.
• A six-week syllabus (one 20-minute session a week, table below) covers all five with activities a 13-year-old will tolerate.
• OECD’s PISA 2025 data links heavy AI use for summarizing and drafting to lower science scores; the association isn’t causal, but it’s the reason to teach “when not to” alongside “how to.”
• The best check for whether it worked isn’t a quiz. It’s whether your teen can explain, unprompted, why they trusted a specific answer.

What the Common Sense survey actually measured

“Teens in the AI Era: Schoolwork and Skills That Matter” is a nationally representative survey of 1,017 U.S. teens ages 13 to 17, released by Common Sense Media on August 18, 2026. It asked teens what they use AI for, how they use what it produces, and what adults have told them about it.

The headline numbers:

• 70% use AI for schoolwork.
• Of those, 63% say they get an answer from the tool. Twenty-five percent use it as-is, 31% rewrite it to sound like themselves, and 35% improve it with their own knowledge (the categories overlap).
• 27% say a teacher has discussed safe AI use with them. Seventy-three percent say no teacher ever has.
• 44% have had an AI tool blocked at school; 59% of those switched to a personal device.
• 37% say they don’t understand their school’s rules on AI.

Two other figures are easy to miss and matter more for parents. Thirty-eight percent of teens say AI leaves them generating fewer ideas of their own, and 39% feel they miss learning opportunities. Teens know something is off. They don’t have the vocabulary or the permission to fix it.

The gap isn’t new. Pew Research Center found in January 2025 that 26% of teens had used ChatGPT for schoolwork, double the 2023 share. Eighteen months later, “a quarter” became “seven in ten.” Instruction did not keep pace.

To teach teens safe AI use, treat it as five habits, not one lecture

Safe AI use is a bundle of specific behaviors, and each one fails in a different way. When a school does cover AI, it usually covers one (plagiarism) and calls it done. Here is the full set.

Verification. A large language model predicts the next likely word based on patterns in its training data. It does not look anything up unless it’s been given a search tool, and even then it can misread what it found. It produces confident text whether or not the underlying fact is true. The habit: never move a factual claim from an AI answer into schoolwork without a second source.

Data protection. What a teen types into a chatbot may be stored, reviewed by humans, or used for training depending on the product and settings. OpenAI’s parental controls page, for instance, offers parents a switch so a linked teen’s conversations “won’t be used to improve models,” which tells you the default without that switch is different. The habit: no full names, addresses, school names, photos of faces, or anything you’d be embarrassed to see printed.

Disclosure. Rules differ by teacher and by assignment. Thirty-seven percent of teens don’t understand their school’s rules. The habit: ask before the assignment, not after the accusation, and keep the chat history as a receipt.

Emotional distance. Common Sense’s July 2025 companion survey found 72% of teens had used an AI companion and a third had chosen an AI over a person for a serious conversation. OpenAI’s August 2026 teen update now instructs ChatGPT to avoid romantic language and to not “imply that it has feelings or consciousness.” That the company had to write that rule tells you the pull is real. The habit: the tool is a tool, and the moment it feels like a friend is the moment to talk to a human.

Knowing when the tool is doing the learning. OECD’s PISA 2025, released September 8, 2026, found 46% of students across the OECD use AI chatbots weekly or more for learning, and that students who use AI mainly to summarize, draft, and research scored lower in science than moderate users. The OECD is careful to call this an association, not a cause. But it’s the reason “how to use AI” has to include “when not to.”

How to Teach Your Kid About Safe AI Use

The six-week plan below is built for 13 and up, because that’s the age most tools require and the age UNESCO recommended as a floor for independent use in its 2023 guidance. Younger kids still need a version. Here’s how it scales.

Ages 5–8: The “computer that guesses” game

Sit together with a voice assistant or a kid-safe chatbot and ask it three questions you already know the answers to, including one it will likely get wrong (a fact about your own family works well). When it errs, celebrate the catch. The lesson is simple and durable: the computer guesses, and guesses can be wrong. No accounts, no independent use.

Ages 9–12: The two-source rule

Pick a homework question with a factual answer. Have your kid ask a chatbot, then find the same fact in a library book, a .gov or .edu page, or an encyclopedia. Do they match? Do this three times over three weeks. Kids this age can also learn the privacy rule as a physical list taped by the computer: no name, no school, no address, no face.

Ages 13+: The six-week syllabus

Run the table below as written. Twenty minutes, once a week, same night. Your teen drives; you ask questions. The sessions build on each other, so keep the order.

The question to ask: “Show me one thing the AI told you this week that you checked, and tell me how you decided whether to trust it.”

The six-week home syllabus, week by week

Each session has one activity and one takeaway. Print the table, stick it on the fridge, and cross weeks off.

Week	Lesson	Activity (20 min)	The habit it builds
1	How the model actually works	Ask the same question three ways; compare the answers; ask the AI for its sources and check whether they exist	Verification; understanding that output is prediction, not lookup
2	What goes in stays in	Read the privacy settings of the tool your teen uses together; find the training-data toggle and the memory toggle; decide what to set	Data protection
3	The rules at your school	Pull the teacher’s syllabus or the district policy; write a one-line summary per class (“Allowed for brainstorming, not drafts”)	Disclosure
4	When it feels like a friend	Read three real chatbot exchanges (yours, not theirs); spot the flattery; talk about why companies had to add “no feelings” rules	Emotional distance
5	Study mode vs. answer mode	Take one homework problem; solve it with AI giving hints only; then solve a similar one with no AI; compare how much stuck	Knowing when the tool is doing the learning
6	Your own policy	Teen writes a one-page personal AI policy: what they use it for, what they don’t, how they cite it; you sign it too	Ownership

A few notes on running it. Week 1 works best with a question about something local or recent, where the model is more likely to hallucinate a source. Week 4 is the one parents skip, and it’s the one that matters most by the data. Week 6 is deliberately last; a policy written before the other five is a lecture in disguise.

What to actually do at home between sessions

The syllabus is the structure. These tactics are the culture.

Use the tool in front of them

Teens copy what adults do more than what adults say. Use a chatbot for something real while your kid is in the room: planning a trip, drafting a tricky email, checking a recipe conversion. Narrate the check: “It says 240 grams. Let me confirm that.” You’ve modeled verification without a lesson.

Ask for the chat, not the grade

When homework comes back, ask to see the conversation that produced it, not just the result. This isn’t surveillance; it’s the same as asking to see the rough draft. Teens who know the chat is part of the work stop treating it as a shortcut. If your teen’s account is linked through parental controls, remember that most tools, including ChatGPT’s, don’t give you transcript access. The chat has to be shared voluntarily, which is the point.

Set the account up together, once

Every major tool has a teen mode now. OpenAI’s ChatGPT for Teens, announced August 18, 2026, routes anyone who states an age of 13 to 17 (or whom its age-prediction model flags) into a stricter experience with Study Mode and “responsible homework reminders.” Google’s Gemini has admin-gated access through Classroom. Sit down for thirty minutes, turn on the study defaults, and agree on the settings. Our guide to safety settings for ChatGPT, Claude, and Gemini walks through each one.

Make “I used AI for this” boring

The goal is a household where “I used ChatGPT to outline this” gets the same reaction as “I used a calculator.” Overreact once and disclosure ends. If you need a framework for the homework conversation specifically, what to do when your kid used ChatGPT for homework covers the first conversation.

What not to do

Don’t install the syllabus as a punishment after an incident. Common Sense’s data shows 59% of teens who hit a block at school simply moved to their own phone; a home version of the same block produces the same result. The plan works because it’s offered before anything goes wrong, framed as a skill, and run with your teen in the driver’s seat.

What to Watch For Over the Next 3 Months

• Week 4: Your teen should be able to name the tool they use most, where its privacy settings are, and one thing they changed. If they can’t, redo Week 2 before moving on.
• Month 2 red flags: Chat histories that are always cleared; homework that reads nothing like their speaking voice; a tool being referred to by name as if it were a person (“she said”).
• Month 3 self-check: Ask for one example from the past week where they chose not to use AI for something, and why. A teen who can answer that has learned the fifth habit. A teen who can’t needs Week 5 again, with a different subject.

Frequently Asked Questions

My teen’s school bans AI. Do I still need to do this?

Yes, arguably more so. Common Sense found that 44% of teens had a tool blocked at school and 59% of those switched to a personal device. A ban moves the use out of sight; it doesn’t remove it. A teen using AI on their own phone with no instruction is the exact population this syllabus is for.

Isn’t 13 too young for any of this?

Thirteen is the minimum age most tools set and the floor UNESCO recommended in 2023 for independent use. Below that, the Ages 5–8 and 9–12 activities above are enough. The point at 13 isn’t to hand over the tool; it’s that your kid likely already has access, and the training should arrive at the same time.

What if I don’t understand AI well enough to teach it?

You don’t need to. The syllabus is built so the teen drives and you ask questions. “How do you know that’s true?” and “Where did that number come from?” are the two questions that do most of the work, and they’re the same ones you’d ask about any source.

How do I know if the AI is doing the learning instead of my kid?

Week 5’s comparison is the cleanest test: one problem with AI hints, one similar problem without. If the second one is a wall, the tool has been carrying the load. OECD’s PISA 2025 found moderate use associated with better outcomes than heavy use; the pattern, not the tool, is what you’re watching.

Should I read my teen’s chats?

Most parental controls, including ChatGPT’s, deliberately don’t let you. Asking to see the chat that produced a piece of homework is reasonable and normal, like asking to see a draft. Reading everything they’ve ever typed is a different thing, and the research on monitoring versus trust suggests it costs more than it protects.

---

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. Common Sense Media. (2026). “Teens in the AI Era: Schoolwork and Skills That Matter.” August 18, 2026. https://www.commonsensemedia.org/research/teens-in-the-ai-era-schoolwork-and-skills-that-matter
2. OECD. (2026). “PISA 2025: Students’ reading and mathematics performance declined sharply across the OECD.” Press release, September 8, 2026. https://www.oecd.org/en/about/news/press-releases/2026/09/pisa-2025-students-reading-and-mathematics-performance-declined-sharply-across-the-oecd.html
3. Pew Research Center. (2025). “About a quarter of U.S. teens have used ChatGPT for schoolwork, double the share in 2023.” January 15, 2025. https://www.pewresearch.org/short-reads/2025/01/15/about-a-quarter-of-us-teens-have-used-chatgpt-for-schoolwork-double-the-share-in-2023/
4. UNESCO. (2023). “UNESCO: Governments must quickly regulate Generative AI in schools.” September 7, 2023. https://www.unesco.org/en/articles/unesco-governments-must-quickly-regulate-generative-ai-schools
5. OpenAI. (2026). “Introducing ChatGPT for Teens: Built for learning, backed by protections.” August 18, 2026. https://openai.com/index/chatgpt-for-teens/
6. OpenAI. (2025, updated 2026). “Introducing parental controls.” September 29, 2025; updated July 13, 2026. https://openai.com/index/introducing-parental-controls/
7. Common Sense Media. (2025). “Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions.” July 16, 2025. https://www.commonsensemedia.org/research/talk-trust-and-trade-offs-how-and-why-teens-use-ai-companions