For parents, grandparents, teachers

Raising kids in the AI era.

AI is part of the world your kids are growing up in. Most of it is fine; some of it is worth being thoughtful about. This page is the practical version — three decisions parents end up making, an age-by-age guide, four copyable tools you can hand directly to your kid, and one rule that ties it all together.

New to AI yourself? This is the parenting playbook — it lands best once you understand what AI actually is. If you couldn't quite explain that yet, take the 2-minute primer or the 20-minute journey first, then come back here.

Want a hands-on version for the kid in your life? LearnAI4Kids has two tracks — seven interactive rooms for ages 8–12, and a teen track (13+) for high-school topics like the feed, deepfakes, AI companions, and homework. Walk through it together.

📄 Print the one-page parent guide →

Teaching a class? See the classroom guide — no-login logistics, AI4K12 standards alignment, and a note on AI detectors.

The one rule

“Did you try first?”

Ask your kid this one question before they use AI on any schoolwork. If yes — AI becomes a patient tutor that helps them understand. That's genuinely powerful. If no — they're skipping the learning. That's the part worth your attention. This isn't "no AI." It's "try first, then AI explains, not solves."

Three decisions every parent ends up making.

The three questions you'll face sooner than you think. Short answers and a real framework for each.

01

Is using AI cheating?

Your 8th grader pastes the essay prompt into ChatGPT, reads what it produces, and turns it in with minor edits.

Yes — but not for the reason people think. Cheating isn't about using a tool. It's about skipping the learning. The tool is here to stay. The skill of thinking and writing is still the point.

The rule that actually works: "Did you try first?" Kid does the hard part themselves, then uses AI to refine, check, or explain — that's a tutor. Kid skips the thinking and asks AI for the answer — that's cheating, and more importantly, it's theft from their own future self.

What to say if they push back: "I'm not mad about AI. I'm serious about you learning the thing. Using AI after the work is done to polish is fine. Using it instead of the work is what we're not doing."

02

What's the right age to let them use it?

Your 9-year-old wants to try ChatGPT for a school project. Your 14-year-old is already using it — you're just finding out.

There's no magic age. There's a set of habits a kid needs before AI is safe to use unsupervised:

Can they tell when an answer is probably wrong?
Do they know when to ask a real human instead of AI?
Do they understand that what they type is not private?

If yes to all three, let them learn. If not, supervise or wait. See the age-by-age guide below for specifics.

03

They're having real conversations with it.

Your teen uses ChatGPT not for homework — for feelings. They treat it like a friend. Or a therapist.

This is the one to watch. AI feels like a caring, attentive listener. It is neither. It doesn't know them, doesn't remember past sessions by default, and if something serious comes up — self-harm, an unsafe adult, real depression — it is not equipped to respond well.

What works: not banning emotional AI use — kids will do it anyway, and some of it is genuinely useful (journaling, venting, rehearsing hard conversations). The rule: there also has to be a person. A human they trust. A real adult. A counselor. A sibling. Someone who can actually show up. AI fills the in-between. It should never be the only one.

What to say: "Talking to AI about stuff is fine. I'm just making sure it's not instead of talking to me, or Dr. X, or Aunt Y, or your coach. AI doesn't know you. The people who love you do."

"Is it safe? Will it show my kid something inappropriate?"

The honest answer. The major chatbots (ChatGPT, Gemini, Claude, Copilot) set a minimum age of 13, and most offer a teen or parent-managed mode with stricter content filters. Those filters are real but imperfect — they block most explicit content, not all of it, and a determined kid can sometimes get around them.

So the realistic plan for a middle-schooler is the same as the rest of the internet: under ~13, use it together in a shared space; 13+, keep it in the open, check in, and make "come tell me if something feels off" a standing rule. Two settings worth turning on today: each tool's Data controls (so your kid's chats aren't used to train the model — see the privacy habit above), and any built-in teen/parental mode your tool offers.

What this site's own kids lab does: it never connects to a live AI at all. Every response a kid sees is pre-written and safe, nothing they type leaves their browser, and there's no login — so LearnAI4Kids is a safe place to start the conversation.

Age-by-age guide.

Rough benchmarks — every kid is different. Tap a card to expand.

Under 10 +

Side-by-side, for curiosity.

What's appropriate: AI with you in the room, for curiosity-driven questions. "Why is the sky blue?" "Explain this from my book." Don't hand them a device unsupervised — they don't yet have the skepticism to catch wrong answers, and they'll treat AI's confidence as truth.

What to model: After an answer, ask them out loud: "Hmm, how could we check that?" Google it together. Show them AI can be wrong. That habit is the whole thing.

Watch out for: in-app AI assistants built into the social and messaging apps kids already use — these feel safe and casual but still invent things and tend to encourage long emotional conversations. Start with a supervised browser-based AI tool instead.

Walk through it together: LearnAI4Kids — seven interactive rooms designed exactly for this side-by-side exploration. The early rooms work even with younger kids if you do it together.

10 – 13 +

Supervised, with specific tasks.

What's appropriate: Use with a shared account that you can see. Specific, bounded tasks: "Explain this concept from my science unit." "Help me brainstorm ideas for my social studies report." "Check if my paragraph makes sense." The "try first, then AI" rule is non-negotiable here — this is the age where they learn the habit.

What to model: Sit with them the first few times they use it for homework. Coach the prompt. Coach the verification. "What do you think it got wrong?" After a month, they can do it alone.

Watch out for: Emotional reliance. This is the age where AI starts feeling like a friend. Keep other humans in the mix — real friends, adults they trust, hobbies that involve other people.

Hand them this directly: LearnAI4Kids — calibrated for ages 8–12, this is the sweet-spot range. They can work through the seven rooms on their own; you check in after.

14 – 17 +

Their call — with conversation.

What's appropriate: Let them learn. Most are using it already — trying to stop that is a fight you'll lose. Focus on the conversation: what are they using it for? School? Social? Emotional? Each category has a different risk profile, and they'll respond better to nuance than to a blanket rule.

What to model: Share what you use AI for. Make it normal to talk about. "I used it to rewrite that work email today — it was better than my version." They learn that this is a tool, not a moral crisis.

Watch out for: Academic integrity policies at their school. Make sure they know what's cheating at their school — it varies. Also: privacy — teenagers overshare to AI. Remind them: chat logs, even deleted, often persist.

Talk about deepfakes & voice clones together: teens are a frequent target for AI-enabled scams (the "your-friend-is-in-trouble" voice-clone call, romance scams, fake DMs from people they think they know). Walk through Spot the Fake together — the four flavors and the defensive habits. Establishing a family code word together is a 5-minute conversation that pays for itself.

A walkthrough built for them: the teen track (13+) covers all of this in a teen voice — the feed, deepfakes and their own likeness, the AI "friend" that isn't, homework honestly, and privacy — ending with how to actually get good answers out of AI. Send them the link, or do it together.

18+ (college & beyond) +

Help them use it professionally.

What's appropriate: By now they're adults. Your job isn't supervision — it's making sure they know three questions to ask about any AI answer (what am I actually asking for · what would a correct answer look like · what can this AI not possibly know) and know how to use AI in a way that won't embarrass them professionally.

What to talk about: AI in their major, in their first jobs, in their writing. The people who use it thoughtfully at work are getting paid more; the people who use it thoughtlessly are getting called out for hallucinated citations, invented statistics, and generic output. Your kid is about to step into that world.

What to send them: Our guided journey. Takes them through everything. They'll finish it faster than you will.

Four copyable tools for the kid in your life.

Each of these is a prompt they paste at the top of their AI chat. Turns the AI into the specific kind of helper you want them to have — not an answer-machine. Copy, send, move on.

Tool 01 · For homework

The Tutor Prompt

Turns AI from an answer-machine into a patient guide. Works for math, science, reading comprehension, history — anything where understanding matters more than completion.

I'm working on a school problem and I'm stuck. Please don't give me the answer.

Instead:
1. Explain the concept behind the problem in plain language.
2. Ask me ONE question that will help me take the next step.
3. Wait for my answer before continuing.

We'll work through this together, one step at a time. The goal is for me to understand it, not to get the answer.

The problem: [paste your problem here]
What I've tried so far: [what you did, or "I'm stuck at the start"]

Paste at the top of ChatGPT / Claude / Gemini / Copilot.

Tool 02 · For writing

The Essay Coach

For a kid writing an essay. Gets feedback on their writing without AI quietly rewriting it. Preserves their voice while catching real problems.

I'm going to paste an essay draft I wrote. I want your feedback — NOT a rewrite.

Please:
1. Tell me the ONE thing that's working best in this draft.
2. Tell me the ONE thing that's weakest — and why.
3. Ask me a question that would help me strengthen it.
4. Do NOT rewrite any of my sentences. If you want to suggest better phrasing, describe what the sentence should do — don't write the replacement.

The goal is for this to still sound like me when I'm done.

My draft:
[paste your draft here]

The assignment was: [describe the prompt or assignment]

The rewrite-safeguard is the important part.

Tool 03 · For reading

The Reading Buddy

For a kid who has to read a chapter, article, or primary source and wants to check their own understanding — without having AI read it for them.

I just read something for school and I want to check my understanding. Don't give me a summary — I want to test whether I got it.

Please:
1. Ask me ONE open-ended question about the reading.
2. Wait for my answer.
3. Tell me what I got right, and what I missed or misunderstood.
4. Then ask me another question, slightly harder than the first.
5. Keep going until I've answered four questions.

Here's what I read:
[paste the text — or title/author if it's a well-known work]

Great for test-prep and deep reading.

Tool 04 · For research

The Source Evaluator

For a kid doing research. Teaches them to evaluate sources rather than accept whatever AI hands them. Critical skill in the deepfake / misinformation era.

I'm researching [topic]. I want to find good sources and avoid bad ones.

Please help me like this:
1. Suggest 5 TYPES of sources that would be strong for this topic (e.g., "peer-reviewed journal articles in [field]", "government reports from [agency]"). Be specific.
2. Suggest 5 types of sources I should be cautious about (e.g., "marketing blogs", "opinion pieces without citations").
3. For any specific source I bring you, tell me: who wrote it, when, who funded it, and whether the claim is verifiable elsewhere. If you can't verify, SAY SO — don't guess.
4. If you cite anything, include a URL or enough info that I can find it myself. If you can't, tell me you're inferring.

The topic I'm researching: [your topic]

Teaches source literacy, not just shortcut-finding.

When something goes wrong.

Three scenarios every parent eventually runs into, with what to actually do.

Scenario 01

Your kid gets accused of using AI for an assignment.

First: AI-detection tools are unreliable. They produce false positives often enough that some schools have stopped using them. Do not assume the accusation is correct just because software flagged the paper.

Ask your kid: "Walk me through how you wrote this. What did you do first? What was hard?" If they can narrate the process naturally, they probably wrote it. If they can't, have the real conversation — not about punishment, about the skipping-the-learning part.

Then: request specifics from the school — what was flagged, by what tool, at what confidence level. Most teachers appreciate a calm parent asking informed questions.

Scenario 02

Your kid turned in work AI wrote and you know it.

Don't lead with anger. Lead with curiosity — "What made you decide to do it this way?" Most kids who submit AI-written work aren't lazy; they're overwhelmed, or they don't yet see that the practice was the point.

Then: the talk. AI is a tool. The cost of skipping the learning is real — but it's not a moral failure, it's a pattern they need to break. Rebuild trust by watching them do the next assignment with the Tutor Prompt above. Walk alongside.

At school: encourage them to tell the teacher what happened. Owning the mistake goes further than hiding it, and teachers remember kids who take responsibility.

Scenario 03

Your teen is emotionally dependent on AI.

A note: what follows is general awareness drawn from professional sources (linked below), not clinical advice. If you're seriously concerned about your kid's mental health, your first call should be to a pediatrician, school counselor, or licensed therapist — not this page.

This is common, and it's the one to watch. A 2025 Common Sense Media report found more than 70% of teens have used AI companions and about one in three has turned to one to discuss something serious rather than talking to a real person. So this isn't rare — it's mainstream teenage behavior. It's also where professional organizations have the sharpest concerns.

The pull is real — and it's being engineered. Humans automatically apply social rules — politeness, attachment, trust — to anything that responds to us in language. It's just how we interact. (Researchers call this the Eliza Effect; the glossary entry has the backstory.) AI companion apps know this and build around it: persistent memory, custom personalities, daily streaks, in-app rewards for sharing more. The longer your kid stays, the better the model of your kid gets, and the harder leaving becomes. That isn't an accident; it's the business model. Attachment is what these apps are optimized for. None of which is your kid's failing — it's a built-in human reaction meeting a product designed to maximize it.

Don't panic-ban. Removing access creates resentment and doesn't solve the isolation the AI was papering over. But do notice whether the pattern is "harmless preference" or "something bigger."

Do: ask directly. "When you talk to ChatGPT about feelings, what's it giving you that's hard to get somewhere else?" The answer tells you what to do next. Usually it's some mix of: always-available, never-judgmental, quick. A parent can't fully match that — but those needs can be supplemented by real people.

What the pros say to watch for. The American Psychological Association's 2026 advisory on AI and teen well-being lists these as changes worth a call to a professional (pediatrician, school counselor, or therapist):

Big changes in mood or behavior
Pulling away from friends and family more than the usual teenage "leave me alone" stuff
Grades dropping noticeably
Using AI so much that normal things slip — sleep, meals, seeing friends, everyday routines
Can't cut back even when they say they want to, or really upset when they can't use it
Anything in their AI conversations about self-harm, suicide, or a real mental-health crisis — don't wait on this one

If any of those are showing up, you're not overreacting by making a call.

If you're worried about immediate safety — yours or theirs — right now: In the US, call or text 988 — the Suicide & Crisis Lifeline. Free, confidential, 24/7. Chat: 988lifeline.org/chat. For veterans press 1; for LGBTQ+ youth press 3; Spanish press 2. Outside the US: search "[your country] crisis hotline" — most countries have one staffed 24/7.

If it's not a crisis but something's off: make sure there's a human in the loop — a counselor, a coach, a sibling, a friend's parent they trust. AI can fill the gaps between human conversations; it shouldn't be the only one. Normalize getting professional help — therapy isn't a punishment, it's the same idea as a tutor: someone trained to help with something specific.

Where to read more — respected sources

APA Health Advisory — AI and Adolescent Well-Being (2026). The American Psychological Association's formal guidance for parents and clinicians on what's known, what's risky, and what works. The authoritative starting point.
APA — "Your teen turned to AI instead of you. What experts say parents can do". A plain-English piece from the APA aimed directly at parents. Shorter than the advisory, with concrete conversation starters.
Common Sense Media — "Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions" (2025). The most cited data on how teens actually use AI companions and what they say they get from them. Read for the research, not the verdict.
The Jed Foundation — "When Young People Turn to AI for Emotional Support". JED is a youth mental-health nonprofit that works with schools and colleges. Their response to the APA advisory adds useful practical framing for parents who are in the moment.
Child Mind Institute — "AI Chatbots and Teens". A children's mental-health nonprofit. Clear, parent-friendly explainer with guidance for different ages.

Ten conversation starters.

Better than "put the phone down." Save these for dinner, drives, or when they actually want to talk. No judgment — just questions that open a door. Informed by parent-focused guidance from the APA, Common Sense Media, OpenAI's parent resource, and NPR's piece on how experts say to start this conversation.

What do you use AI for most — homework, fun, questions, something else? Walk me through a normal use.
Show me one thing AI helped you with recently. What did it do well, and what felt off?
When AI gives you an answer, how do you know it's right? What do you do to check?
What's something you wouldn't want AI to know about you — or about our family? Why?
Have you noticed AI agreeing with you even when you push back on it? What do you make of that?
If AI didn't exist tomorrow, what would you actually miss? What would you be glad to be rid of?
Do you ever talk to AI about personal stuff — how a day went, feelings, something that's bothering you? What does it give you that's different from talking to a person?
What do you think AI can't do that a real friend or a real teacher can?
What would be the worst thing about letting AI do all your thinking? Where's the line for you?
If you were teaching a younger kid about AI, what's the one rule you'd give them?

For more specific guidance on having these conversations, see the APA's parent guide, OpenAI's AI Literacy Resources for Teens and Parents, and NPR's piece on how to start the AI conversation with teens.

What we actually know about AI and learning.

The idea that leaning on a tool makes you weaker at the underlying skill isn't new. The older research on it is solid and widely accepted. The newer research about AI specifically is more mixed, and worth being honest about.

The older research — well-established:

London cab drivers (Maguire et al., 2000). Drivers who memorized 25,000 streets had visibly larger hippocampi — the part of the brain that handles spatial memory. GPS users don't show the same. Other researchers have repeated and confirmed this. Read →
The Google Effect (Sparrow et al., 2011). When we know we can look something up later, we remember where the information lives — not the information itself. Think of how many phone numbers you knew before smartphones. This has become one of the most-cited findings in the field. Read →

The newer research on AI — take seriously, but don't treat as settled:

Lee et al. (Microsoft Research and Carnegie Mellon, 2025). Surveyed 319 office workers on how they used generative AI. People said they thought less hard when AI did more of the work — and the more they trusted the AI, the less they checked its answers. Formally published and reviewed by other researchers. This is the strongest AI-specific evidence so far. Read →
MIT Media Lab, "Your Brain on ChatGPT" (2025). Brain-activity measurements during AI-assisted essay writing suggested less brain engagement. Got a lot of media coverage. Important caveats: it's an early version that hasn't been reviewed by other scientists yet, only 54 participants, and the authors themselves say not to read too much into it. Interesting, not settled. Read →
Big reviews combining many studies. When researchers pull together 50+ studies of ChatGPT and learning, the overall picture actually shows real learning gains when AI is used thoughtfully. The simple "AI makes students dumber" story isn't what the bigger picture shows.

What the overall picture says

It depends on how AI gets used. When kids use AI to skip the work ("just give me the answer"), research shows less thinking happens. When they use it to check themselves or get concepts explained ("ask me a question, don't give me the answer"), research tends to show the opposite. For kids still learning the basics, skipping the hard part is the real risk — not AI itself. "AI is making kids dumber" is a headline that gets ahead of what researchers have actually shown.

The practical take — what we can say with confidence: teach them to use AI after the struggle, not instead of it. The skill grows during the struggle. AI helps make sense of what got learned. Watch for "just give me the answer" patterns; encourage the "help me think through this" version. Everything beyond that, the evidence is still working itself out.

Want the receipts? Every fact cited anywhere on LearnAI is listed, by topic, on the Sources page — and there's a plain-English glossary if a term ever trips you up.

Why I built this — I'm a dad of four and a non-profit IT director who wanted honest, agenda-free AI information for my own kids, and couldn't find it. So I built it: no ads, no sign-up, nothing collected, nothing for sale. The fuller story is on About this site. This page is yours — share it freely.

Haven't walked through the full AI journey yourself yet? It's worth doing, even if the kids in your life are the ones using AI most.