AI for your family: kids, teens and seniors

Time: about 25 min reading + 60 min for the family conversation (the weekend project is extra)

⚠️ About age limits. Service rules change, and each service sets its own minimum age. As of October 2026: under Anthropic's rules, Claude (claude.ai and the app) is for adults 18 and older only; ChatGPT allows users from age 13 (with a parent's permission until 18) and offers parental controls through linked accounts. Before you give a child any AI service, open its rules and check the minimum age. This lesson offers general guidance, not medical, legal or psychological advice.

The gist

AI showed up in your home before you noticed. Your daughter asks Snapchat's My AI about a boy in her class. Your son gets his homework done with ChatGPT in two minutes. Mom wants a recipe for whatever's left in the fridge. Dad is trying to work out how to explain to the kids why grass is green, and he can't remember himself.

You can't ban it; that's like banning the internet in 1995. You can't give total freedom either: a child can't tell a confident AI answer from a correct one. You need a third way: rules, the right tools for each age, and teaching everyone to check.

In this lesson we'll look at how to build good AI habits for the whole family, from a five-year-old who listens to AI-made bedtime stories to a grandmother who uses AI to make sense of the instructions that came with her medication. Is ChatGPT safe for kids? How do you get a grandparent started with AI in easy steps? We'll cover both. No panic, no hype, just practical steps.

🎨 Picture this: AI is a set of tools in Dad's garage. The screwdriver sits on a low shelf; a 5-year-old can use it with a grown-up watching. The power drill: a teenager handles it alone, once Dad has shown how to hold it. The table saw: adults only, with safety glasses. Same garage, different rules for who gets to use what. Lock the garage for good, and the kids never learn how to use the tools.

🎯 Decision tree: which AI for which age

The real question isn't "should we allow AI?" It's "which tool, and which rules, for this particular age?"

Ages 5-10: supervised only

• ✓ AI is used together with a parent (never alone)
• ✓ Educational tools made for kids only (for example, Khanmigo, which families can get in the US)
• ✓ No social AI (Snapchat's My AI restricted)
• ✓ Limited time (like TV)

Ages 11-15: partial independence

• ✓ From 13, can use it alone, with a weekly check-in with a parent
• ✓ Learns the difference between "AI helps me learn" and "AI does it for me"
• ✓ Parental controls turned on
• ✓ Family rules written down and signed

Ages 16-18: full access + critical thinking

• ✓ Their own account on a service that allows their age (ChatGPT: ages 13-17 with parental controls; Claude: 18 and up only)
• ✓ The main lesson: checking AI for mistakes
• ✓ Privacy education: what not to share with AI
• ✓ Getting ready for adult life with AI

Older parents and grandparents: voice mode + scam education

• ✓ Simple interfaces (voice)
• ✓ Learning to recognize AI-powered scams
• ✓ No medical or financial decisions based on AI alone

🎨 Picture this: a tricycle → a bike with training wheels → a regular two-wheeler. Each stage comes with its own level of freedom and its own risks. Take the training wheels off too soon, and they fall. Leave them on too long, and they never learn to balance.

Key concepts

• Supervision tier: how much independence a child has with AI. Full supervision (a parent sits right there), checkpoint supervision (a weekly review), trust but verify (full access, with occasional check-ins)
• Parental controls: limits built into AI tools, such as age checks, content filters, schedules and switches for individual features; kids' learning tools may also show the parent the chat history
• AI as a tutor vs. AI as a ghostwriter: the critical difference. Either AI explains how to solve it (you learn), or it solves it for you (you don't)
• Hallucinations: AI confidently making things up. It can give a wrong answer with total confidence, and kids can't spot this unless someone teaches them
• Privacy boundaries: what never goes into AI: names, documents, medical records, finances, photos of children
• AI-enabled scams: fraud that uses AI (cloned voices, fake video calls); especially dangerous for older adults
• Family AI agreement: a written agreement about which AI rules your family has decided on
• Voice-first interfaces: talking to AI out loud instead of typing; essential for older adults and for kids who can't write yet

Theory

3 age groups, 3 approaches

Ages 5-10: AI as play and explanations, ALWAYS supervised

At this age, AI shouldn't be alone with a child. Not because "AI is bad," but because a child doesn't yet understand that AI can be confidently wrong. If a six-year-old asks "Is it true Santa flies faster than an airplane?" and AI says "Yes!" in a confident voice, the child will believe it.

What works at this age:

• AI helps make up bedtime stories (the parent generates the text in their own account and reads it aloud)
• AI answers kids' questions ("Why is the sky blue?"), with a parent nearby
• AI draws pictures from the child's description (Midjourney, with a parent)
• Khanmigo from Khan Academy: an educational AI built with child safety in mind. A parent can see the chat history and gets alerts. Note that the family version is sold only in the US (as of October 2026); elsewhere, at this age, use an assistant from your own account with your child sitting next to you

What doesn't work:

• Snapchat's My AI without parental controls: a child asks about relationships, and AI answers as if it were talking to an adult
• Regular general-purpose chatbots (ChatGPT and others): officially not meant for anyone under 13; the services' own rules don't allow it
• ChatGPT's voice mode: a child can talk to it for hours with no one supervising

A time guideline: no more than 30 minutes a day of talking with AI at this age (that's our suggestion, not an official standard). It's not an electronic babysitter; it's a tool for exploring the world.

🎨 Picture this: AI at this age is like an encyclopedia at the public library. The child can page through it alone, but a grown-up sits nearby to explain whatever's confusing and to set things straight if the encyclopedia is out of date.

Ages 11-15: AI as a tutor and a creative tool

This is when kids start using AI on their own, and lots of them do. You won't be able to ban it; their classmates have it. A parent's job isn't to forbid it, but to teach the difference between "AI teaches me" and "AI does it for me."

The main lesson: AI as a helper vs. AI as a substitute.

Rules for this age:

• From 13, their own account (not a parent's) on a service that allows it: they learn to take responsibility for what they do. At 11 or 12, only together with an adult: the parent works in their own account and the child sits alongside
• Weekly review: once a week, the teen shows a parent what they asked AI and talks it through (with the teen's agreement, never in secret)
• No turning in AI-written work as your own: that's the line on academic honesty
• Teach fact-checking: "AI said X? Check it on Wikipedia or with a Google search before you use it"

What to allow:

• Homework help as explanation (not as a ready-made answer)
• Learning languages by having conversations with AI
• Creative work (writing a poem or making up a story together with AI)
• Coding (ChatGPT as a mentor, from age 13; Claude only from 18)

What to block:

• AI companions, meaning chatbots built for friendship and relationships (Replika, Character.AI). Common Sense Media's 2025 risk assessment found they pose unacceptable risks for anyone under 18: emotional attachment to an AI can take the place of real relationships. Since November 2025, Character.AI itself has shut off open-ended chats for users under 18
• Unmoderated chats with AI bots on Discord servers
• Generative AI used to make deepfakes of classmates (that's bullying)

🎨 Picture this: AI at this age is a bike with training wheels. The teen is riding on their own, but if they fall, it won't be a hard fall. A parent walks alongside for the first month, then lets go.

Ages 16-18: AI for productivity and as a critical-thinking partner

At this age, your kid is getting ready for an adult life where AI is like electricity: it's everywhere. The main job is to teach them to check AI, because in college and at work they'll use it every day.

The main lessons at this age:

1. A verification mindset: "AI isn't the source of truth; it's an assistant." Any important fact gets checked against the original source
2. Writing good prompts: how to phrase a request so you get a good answer (people call this "prompt engineering")
3. Choosing the right tool: Claude for long texts (from 18), ChatGPT for general use, Perplexity for research, specialized tools for specific fields
4. Privacy habits: what NOT to give AI: passwords, ID numbers, medical records, other people's personal information

What you can give them at this age:

• A paid plan for an AI assistant (for example, ChatGPT Plus, $20/month in the US as of October 2026), but only if they keep hitting the free plan's limits; Claude Pro once they turn 18
• Subscriptions to specialized tools (Perplexity for research, Cursor for coding), if they actually need them
• An introduction to the API, the way programs connect to AI directly (for the technically inclined)

A family conversation you shouldn't skip:

• Depending on AI too much: warning signs and red flags
• AI and mental health: no therapy through AI alone; reach out to real people
• AI and big decisions: choosing a college, a career. AI helps with the analysis; the decision is yours

🎨 Picture this: AI at this age is like a teen moving from a learner's permit to a full driver's license. They're behind the wheel on their own now, and the car has real power. The rules of the road matter more than ever.

Top 10 ways AI helps families

Specific situations where AI actually helps in family life. Not "transforming your life" (we skip the self-help talk), but getting routine tasks done faster.

1. Homework help (explaining, not doing)

   • Prompt: "Explain to a 10-year-old why 3/4 is bigger than 2/3, using a pizza example"
   • The rule: AI explains → the child solves it on their own

2. Learning languages (conversations with AI)

   • An assistant in voice mode for speaking practice (Claude or ChatGPT work for adults; for kids, follow each service's age rules)
   • Duolingo Max (a paid subscription bought in the Duolingo mobile app; its AI features aren't available in every course): structured lessons plus conversation practice with AI
   • Prompt: "Have a conversation with me in Spanish at a beginner (A2) level about buying bread at a bakery. Correct my mistakes."

3. Creative projects (drawing, music, stories)

   • Bedtime story generator: "Make up a 10-minute bedtime story about a girl named Emma and a talking cat, set in Ecuador"
   • Midjourney (a paid service) draws the picture, and the child colors in the final version on paper

4. Explaining hard things (how the internet works, why grass is green)

   • "Explain electricity to a 7-year-old using a simple everyday comparison"
   • When a parent doesn't remember either, AI helps the parent explain

5. Everyday problems (what to cook with what's in the fridge)

   • "I have chicken, rice, carrots, an onion and tomatoes. Suggest 3 dinners I can make in 30 minutes"
   • A weekly family meal planner

6. Family finances (budget, savings goals)

   • "We're a family of 4. Our income is $X and our fixed expenses are $Y. How much do we need to set aside to save $20,000 for a car in 3 years?"
   • AI does the math (double-check it with a calculator); the family makes the decision

7. Health (nutrition, exercise; NOT medical diagnoses)

   • "Make a home workout plan, 3 times a week, 30 minutes each, for a 40-year-old man, no equipment"
   • If you have a chronic condition or haven't exercised in a long time, show the plan to your doctor first
   • ❌ NOT "My stomach has hurt for two days, what is it?" That's a question for a doctor

8. Travel (planning a trip)

   • "Family of 2 adults + 2 kids (ages 8 and 12), one week in the Great Smoky Mountains, budget $2,000. Plan a route with kid-friendly activities"
   • AI makes a draft → the parents adjust it (check prices, schedules and opening hours on official sites)

9. Paperwork (letters and requests)

   • A letter to the school about transferring your child
   • A request to your bank for a statement
   • A letter to your landlord or HOA
   • Leave ID numbers and your address out of the request: add them to the finished letter yourself

10. Memories (help writing letters to older relatives)

    • "Help me organize a story about Grandpa for a family book. Here's what I remember: [facts]"
    • AI as a ghostwriter for family memories

Safety rules (non-negotiable)

Write these rules down as a family and revisit them every 6 months.

NEVER give AI:

• Family members' full names together with their contact details
• ID numbers (Social Security number, driver's license or passport number)
• Bank account details or card numbers
• Medical information about family members (even your own)
• Photos of your children (especially for generating new images)
• Your home address
• Your child's school and grade together with their name
• Passwords (not even temporary ones)

ALWAYS check with a professional:

• Medical advice → a doctor
• Legal questions → a lawyer
• Big financial decisions → an accountant or a financial advisor
• Emotional or mental health problems → a licensed therapist or counselor (not an AI companion)

Technical rules:

• Kids under 13 don't use AI alone (supervision tier: "full")
• Turn on parental controls wherever they exist: Snapchat (in Family Center, a parent can turn off My AI), ChatGPT (linked parent and teen accounts). Claude has no parental mode: the service is for adults only
• Keep chat history for at least 30 days for the weekly review
• Teach your kids: "AI can be confidently wrong" (hallucinations)

🎨 Picture this: AI is like a smart stranger at a coffee shop. He knows a lot and loves to chat. But you wouldn't hand him your driver's license, show him your bank card or tell him where your daughter lives. Same with AI.

Tools for families (ranked, with prices)

By budget:

• Minimum ($0/month): a free assistant for the parents + supervision for the kids
• Middle: Khanmigo for the kids (in the US) + a paid assistant plan for a parent, if needed
• Full: Khanmigo + ChatGPT Plus + Duolingo Max (add up the current prices on each website before you commit)

AI and school (common situations)

Situation 1: "The teacher is against AI"

• What the teacher says: "If I find out AI wrote this, it's a zero."
• What to do: use AI as a brainstorming partner, not as the writer. A prompt like "Give me 5 essay ideas about The Great Gatsby" is fine. "Write my essay" is not. The final text is always written by your child, in their own words. Also read the class syllabus or the school's academic honesty policy: that's where you'll find what counts as allowed help.

Situation 2: "It's too easy"

• The problem: AI does the homework in 2 minutes → kids don't learn
• The fix: teach them not how to game the system, but how to use AI to understand things more deeply. Your son does the homework himself, then asks AI: "Check my solution and explain where I went wrong." That's how he learns to work with feedback.

Situation 3: "The digital divide"

• The problem: kids from well-off families have AI tutors; other kids don't
• The fix: use free tools. Khanmigo is free for teachers, and some schools have extra programs (check the website). MagicSchool AI has a free plan for teachers. AI assistants have free plans too, but each one has its own minimum age.

Your family's position:

• School teaches the foundation (reading, writing, thinking)
• AI is a tool for after that foundation is in place
• First your child learns to multiply on their own, then the calculator. First your child learns to write an essay, then the AI helper.

AI for older parents and grandparents, in easy steps

Older family members are a category of their own. They're not used to typing long requests. On the other hand, they often have more free time, and real needs where AI helps.

The main problem: complicated interfaces. The fix: voice modes.

Uses that work:

• Explaining medical terms (but NOT diagnoses!): "What does 'hypertension' mean, in plain words?"
• Getting ready for a doctor's appointment: "Help me write down 5 questions to ask my doctor about my trouble sleeping." AI helps prepare the questions; the answers come from the doctor
• Making sense of Medicare jargon: "What's the difference between Medicare Part A and Part B, in plain English?" For anything about your own coverage, call Medicare or your plan directly
• Company when they're lonely (Pi from Inflection)
• Help with instructions: "Explain how to use this TV remote"
• Reading small print (the phone camera + ChatGPT's voice mode)
• Translating documents from another language (cover the name, address and ID numbers first)
• Reminders: "Remind me to take my blood pressure pill at 8 a.m." Only if the assistant you chose actually has reminders and you've checked that one really arrives. For medication, a phone alarm or the phone's own Reminders app is more dependable

Dangers:

• AI scams aimed at older adults. One of the biggest threats right now is the cloned voice. Scammers clone the voice of a grandchild, daughter or son (a short clip from social media is enough) and call an older relative: "I've been in an accident, I need $5,000 wired right away." It's the old grandparent scam, now with a voice that sounds real.
• Fake calls "from the IRS," Medicare or Social Security. The caller demands money or your numbers right now, often by gift card, wire transfer or crypto, and tells you not to hang up or not to tell anyone. Treat that kind of pressure as a red flag.
• The family rule (non-negotiable): before sending any money, ALWAYS hang up and call back on a number you already know. Set up a family code word ("pineapple") that a scammer wouldn't know.

Tools for older adults:

• Pi (Inflection AI): https://pi.ai. A friendly conversation partner with a gentle voice and a simple interface. Free as of October 2026; under the service's terms, it's for adults only.
• ChatGPT voice mode: natural conversation; availability and limits depend on the plan
• Voice modes in other assistants: check what your app offers

🎨 Picture this: AI for Grandma is like a phone with big buttons. She doesn't need 50 features; she needs the 5 that work every time. Voice instead of typing, simple instead of complicated.

A family AI project (creative, week 1)

The best way to teach your family to use AI well is to do a project together. Not "learn AI," but "build something useful with AI."

4 projects to choose from:

1. Family Book AI helps you write a book about your family's history. You interview older relatives and record the conversations, turn the recordings into text with a transcription app (many of them run on OpenAI's Whisper model), and AI organizes the text into chapters. Take ID numbers, addresses and phone numbers out of the text first. You print one copy and give it to the family.

• Time: 1-2 months
• Cost: printing one book (check the price at the printing service you pick)
• Tools: Claude or ChatGPT + a transcription app

2. Children's Storybook AI writes a story with your child as the hero, and Midjourney (a paid service; any other image generator works too) draws the illustrations. Don't upload photos of your child to the generator: describe the hero in words. You print it as a hardcover children's book.

• Time: a weekend
• Cost: printing (check the price at the printing service)
• Tools: Claude + Midjourney + a photo-book printing service such as Shutterfly

3. Bedtime stories on demand Every evening, Claude (in your own account) generates a new story with your child's favorite characters. You save the best ones in an "Our stories" file.

• Time: 10 minutes each evening
• Cost: $0
• Tools: Claude Free

4. Family Recipe Book AI organizes the family recipes Mom cooks from memory. You interview Mom out loud, and AI turns her answers into recipes in book format.

• Time: 1 month
• Cost: printing (check the price at the printing service)
• Tools: Claude + a transcription app + Canva for the design

What matters in a family project: AI handles the routine (organizing, formatting, the first draft), and the family brings the meaning (choosing the stories, checking the facts, the feelings).

Common mistakes (for parents)

❌ Banning AI completely: like banning the internet in 1995. Your child will use it at friends' houses, in secret, without ever learning the safety rules. Worse than measured access.

❌ Unlimited access with no rules: without supervision, a child can spend hours with AI companions, grow attached to them and spend less time with people.

❌ Not teaching kids to fact-check AI: the child learns that "AI is always right," then in college and at work makes factual mistakes that could have been caught.

❌ Using AI as a babysitter: Snapchat's My AI keeps a child busy for hours so the parents get some quiet. It's the TV babysitter of the '90s all over again.

❌ Trusting AI with diagnoses or advice on serious matters: medicine, law, mental health. AI makes confident mistakes. A doctor has years of medical school and residency, plus experience. AI is the doctor's assistant, not a replacement.

❌ Not explaining privacy to your kids: a child tells AI about the family's finances, the parents' problems, fights at school with real names. That data may be used to train models, or it may leak.

❌ Secretly reading your child's chat history: it breaks trust. Better: an open weekly review with your teen's agreement. "Show me what you asked AI about this week" is a conversation, not surveillance.

When to step in (red flags)

Signs that AI use has become a problem:

• Your child uses AI instead of talking to people. An AI companion (Replika, Character.AI) has appeared, and your teen is "friends" with it instead of real friends. An emotional attachment to AI.
• They do 100% of their homework with AI and learn nothing. Grades drop on in-class tests where AI isn't available. The work they turn in doesn't match what they can explain out loud.
• They get mental health advice only from AI. Your teen says "AI understands me better than my therapist does." They spend hours talking through dark, depressing thoughts with ChatGPT.
• They can't explain how they got the answer. Your son turns in a physics assignment but can't walk you through the solution in his own words. That's a sign AI is doing the work for him.
• They hide their AI conversations: they slam the laptop shut when you walk in or refuse to show their chat history. That's normal for a teenager on its own, but together with other red flags, it's time to talk.
• It's affecting their sleep. Your teen stays up at night on Character.AI or other AI chats and is exhausted in the morning.

What to do when you see a red flag:

1. Don't panic, and don't ban it overnight
2. Talk without accusing: "I noticed X. Tell me what's going on."
3. If needed, get help from a professional (a child therapist or counselor). If you're ever worried about your child's safety right now, call or text 988, the Suicide & Crisis Lifeline
4. Revisit the family AI rules together
5. Possibly a temporary break (1-2 weeks without AI)

Recommendations by age

Practice

Step 1: A family conversation about AI (1 hour)

Sit down with your spouse or partner and your kids (if they're 11 or older) for an hour-long conversation with no distractions. The goal isn't to "teach the rules"; it's to agree on them together.

How to structure the conversation:

1. What do we know about AI? (15 min)
   • Everyone says what they've used
   • What they like, what worries them
   • No judging ("that's bad," "that's dumb")
2. What risks do we see? (15 min)
   • School and AI
   • Privacy
   • Depending on AI too much
   • Money
3. What rules do we need? (20 min)
   • Age and access
   • What never goes into AI
   • When parents step in
   • Our family's red flags
4. What do we write down? (10 min)
   • One page titled "Our family AI rules" (on paper or in a notes app)
   • Everyone who took part signs it
   • We review it again in 6 months

Step 2: Setting up parental controls

Snapchat (if your teen uses it):

1. In your own Snapchat app, open your profile and tap the gear icon (Settings)
2. Choose Family Center and send your teen an invitation
3. Your teen has to tap Accept
4. In Family Center, scroll down to Settings, expand it and turn on the "Disable My AI" toggle

Where else to turn on parental controls:

• Snapchat: more for parents at https://parents.snapchat.com
• ChatGPT (ages 13-17): link the parent's and the teen's accounts under Settings → Parental controls. Either side can send the invitation, and the other has to accept. The parent can set quiet hours, turn off voice, memory and image generation, and turn off model training on the teen's conversations. The parent doesn't see what the chats say, so the weekly review below is built on conversation. A teen can also unlink the accounts; if that happens, the parent gets a notification
• Claude: under Anthropic's rules, users must be 18 or older; you can't create an account for a child
• YouTube AI features: YouTube Kids for younger children, Restricted Mode for older ones

Step 3: Start a family AI project (a weekend)

Pick one of the 4 projects (Family Book / Storybook / Bedtime Stories / Recipe Book). Start this weekend.

Example: a children's storybook in one weekend

Saturday morning (1 hour):

• Sit down with your child and pick the main character (your child)
• Pick a theme (an adventure, friendship, a rescue)
• Ask Claude (in your account):

Write a children's story for an 8-year-old about a boy named Jake who finds a talking dog, and together they save the park. 10 pages, 50 words each. Style: kind, with some humor.

Saturday evening (1 hour):

• Read the story with your child
• Revise it together with AI ("make chapter 3 funnier")
• Finalize the text

Sunday morning (2 hours):

• Illustrations: Midjourney (paid) or another image generator. A prompt for each page:

Children's book illustration, watercolor style, a boy named Jake finds a talking dog in a park, warm colors, [describe the scene]

• Pick the 10 best illustrations

Sunday evening (1 hour):

• Put it together as a PDF in Canva (it has a free plan)
• Order a printed copy from a photo-book service such as Shutterfly; check the price and delivery time on its site

Your child sees that AI helped make something real. Not "AI is bad," not "AI is magic": a tool.

Step 4: Setting up voice AI for an older relative

If you have older parents, help them get started with Pi (it's free).

1. On your parent's phone, go to https://pi.ai
2. Pi will ask right away what to call you. To keep the conversations, create an account in your parent's name: tap "Sign up for free" (you'll need a phone number)
3. Turn on voice mode
4. Show them how to start a conversation (just talk)

The first 5 things to show them:

1. "Tell me how to make a good chicken soup"
2. "Explain what the word 'streaming' means"
3. "Help me get my questions ready for the doctor about my blood pressure"
4. "Tell me an interesting fact about the Grand Canyon"
5. "Help me write a letter to my grandson"

What you must explain to your parent:

• AI can make mistakes: double-check anything important
• Never give AI your Social Security number, Medicare number, bank details or passwords
• If someone calls with a relative's voice asking for money, ALWAYS hang up and call back on a number you know
• The family code word: ask for it whenever you're not sure
• A call "from the IRS" or "from Medicare" asking for money or your numbers: hang up and call the official number yourself
• Medication and health: AI explains the words and helps prepare questions; the doctor decides

Step 5: A weekly review with your teen

Every Sunday evening, a 20-minute conversation with your teen about what they did with AI that week.

How the weekly review goes:

1. "What did you do with AI this week?" (5 min)
   • Curiosity, not control
   • Your teen tells you in their own words
2. "Show me 2 or 3 interesting chats" (5 min)
   • Your teen picks what to show
   • You don't go through their history in secret
3. "Did AI tell you anything you're not sure is true?" (5 min)
   • Building a verification mindset
   • "Let's check it together"
4. "Anything you want to tell me?" (5 min)
   • An open question
   • Sometimes important topics come up

If your teen doesn't want to talk, don't push. But do it at least once a month.

Tools and resources

• Khanmigo (Khan Academy): an educational AI for school-age kids; the family version is paid and sold only in the US (check the website for the current price)
• Common Sense Media AI guide: reviews of AI tools for kids, with safety ratings
• MagicSchool AI: tools for teachers, with a free plan
• Pi by Inflection: a friendly conversation partner with a voice mode, a good fit for older adults (adults only)
• AAP guidance on children & tech: recommendations from the American Academy of Pediatrics
• Snapchat Family Center: parental controls for Snapchat, including My AI
• ChatGPT parental controls: how to link a parent's and a teen's accounts
• Duolingo Max: conversation practice with AI for language learners (a paid subscription in the mobile app)
• Minimum age for Claude: Anthropic's rules (18+)
• Anthropic Usage Policy: the rules for using Claude

Key takeaways

You can't ban AI for kids; that's like banning the internet in 1995. You can't give total freedom either; a child can't tell a confident AI answer from a correct one. The third way: rules by age, tools matched to each level, and teaching kids to check. Under 13, only together with an adult (in the US, Khanmigo is built for this), ChatGPT with parental controls for ages 13-17, and Claude for adults only (18 and up).

The main lesson for a teen is telling "AI helps me learn" apart from "AI does it for me." AI explains the derivative → you solve the problem yourself → you learn. AI solves the equation → you copy it → you don't learn. Your family rules on this should be written down and signed.

A family AI project (a book, a story, a recipe collection) teaches good habits better than any lecture. Your child sees that AI helped make something real. Not magic and not evil: a tool in Dad's garage. Pick one project for this weekend and get started.

Older parents are a category of their own, with two priorities: simple voice interfaces (for example, Pi, which is free) and learning to recognize AI scams (cloned voices). A family code word is a must. Before sending money, ALWAYS call back on a number you know.

Checklist

• Had the family conversation about AI (everyone took part)
• Wrote and signed your family AI rules
• Turned on parental controls wherever they exist (Snapchat / ChatGPT)
• Your kids understand that AI can be confidently wrong (hallucinations)
• Posted the "what never goes into AI" list on the fridge
• Started a family AI project (a book / a story / recipes)
• If you have older parents: set up a voice assistant and talked through scams with them
• Picked a family code word
• Scheduled the weekly review with your teen (if you have teens)
• Put a rules review 6 months from now on the calendar

Next lesson

→ How to write a good prompt: the first lesson of the second module, on how to give AI a task so you get the answer you need.

If you landed here without the beginning, go back to Your first 30 minutes with AI.