AI for Families | Frameworks, House Rules, Prompts, Research for Parents

Start Here

The Framework

The AI Framework for Parents and Kids is an eight-module course, free and self-paced, built from the strongest published guidance on children and AI and from years of watching what happens when families wing it. It answers the three questions every parent carries: what is this thing, what should my child be doing with it at this age, and what rules will hold in my house.

Free Course • 8 Modules

The AI Framework for Parents and Kids

Parent mindset, AI literacy at home, the age-band ladder, safety and wellbeing, privacy, school and homework, the household agreement, and keeping the conversation going as your child grows.

Open the course →

Article • The Thinking

The Framework, Explained

The article version: three questions that settle most family AI decisions, household rules you can agree tonight, and the safety and privacy guardrails, with the reasoning behind each one.

Read the article →

Source Library

The Guidance It Draws On

Every source behind the framework, from UNESCO's age guidance to Common Sense Media's product reviews, collected with notes on what each is good for. Check my working; I would.

Browse the library →

The Rules That Hold

Eight House Rules for AI

Rules a family can actually keep, written to survive contact with a clever twelve-year-old. Agree them out loud, put them where everyone can see them, and apply them to the adults too, because your child is watching how you use AI more closely than they are listening to what you say about it.

AI helps you think, it does not think for you.

The house test for any AI use: could you explain the result in your own words afterwards? If yes, it was a tool. If no, it was a substitute, and substitutes are not allowed on school work.

Real name, real address, real school: never.

Nothing that identifies your child, your home, or your family goes into an AI tool. Practise the reflex together until it is automatic, the same way you taught crossing the road.

AI use happens where adults can see it.

Shared spaces, visible screens, and accounts a parent can review. Privacy from siblings is fine; privacy from parents is not, at least until the track record earns it.

The age floors are our floors.

Most AI tools set a minimum of 13, several 18. In this house the published minimum is the household minimum, and a tool that hides its minimum does not get used at all.

We check what the machine tells us.

AI states wrong things with total confidence. Make verification a family sport: when the chatbot answers, someone checks a second source, and whoever catches an error wins the point.

AI is not a friend.

It is software that predicts words. A child can enjoy it, use it, even laugh with it, but the moment it becomes a confidant, we talk. Companion chatbots marketed as friendship are off the table entirely.

Tell us, and nothing bad happens.

If something an AI said scared, confused, or embarrassed them, they can bring it to you and the response will be help, not punishment. Lose this rule and you lose all visibility into the other seven.

The rules get renegotiated, on schedule.

Sit down each school term and revise the agreement together as your child matures and the tools change. A rule set that never updates gets quietly abandoned instead.

A Caribbean grandmother and two grandchildren sharing a tablet together on a veranda in golden afternoon light — The rules work best when the whole household keeps them, grandparents included.

Match the Tool to the Child

The Age-Band Ladder

The most common mistake families make is treating AI as one decision instead of a ladder a child climbs. What is right at fifteen is wrong at eight. The full ladder, with activities and conversation scripts for each band, is Module 3 of the framework; this is the short version.

Under 8Together only

No direct AI use. A parent may use AI with the child watching: generating a bedtime story together, asking it silly questions, and narrating out loud what the machine is and is not doing. The lesson at this age is that it is a machine.

8 to 12Supervised & scaffolded

Shared sessions on a parent's account, with the parent present. Creative play, curiosity questions, and the first verification habits. The child drives, the parent sits beside, and sessions end with "what did it get wrong?"

13 to 15Monitored independence

Own use of age-appropriate tools under the house rules: visible screens, no personal data, disclosure on schoolwork. This is the band where the homework shortcut temptation peaks, so the thinking-first rule earns its keep.

16 and upCoached autonomy

Near-adult use with parents as advisers rather than supervisors. The conversation shifts to judgement: bias, persuasion, deepfakes, what AI means for the careers they are choosing, and how to use it as an edge rather than a crutch.

A Caribbean father and teenage son talking on the front steps of a turquoise island house at golden hour, a laptop open between them — By the top of the ladder, the goal is conversation, not supervision.

Working Tools

The Parent Prompt Library

Eight prompts that turn AI into the family assistant it should be: one that asks your child questions, feeds their curiosity, and hands the thinking back. Copy one into the tool your family uses, replace the bracketed parts, and stay in the room while it runs.

Homework Coach, Not Homework MachineHelp that never writes the answer.

You are a patient homework coach for my [age]-year-old working on [subject/topic]. Never give the answer or write any part of their work. Guide them with one small question at a time, wait for their reply, and praise the reasoning, not the result. If they are stuck three times in a row, give a hint, not a solution. Start by asking them to explain the problem in their own words.

Bedtime Story Co-WriterThe child invents, the machine assists.

We are writing a bedtime story together. My child is [age] and wants a story about [their idea]. Write only two sentences at a time, then stop and ask my child what happens next. Weave their answer in, keep the language warm and age-appropriate, and let their ideas steer the plot even when they are wonderfully ridiculous. End the story within ten turns.

The Why MachineFor the question phase, instead of "go ask the tablet".

My [age]-year-old asked: "[their question]". Explain the answer at their level in under 100 words, then ask them one follow-up question that makes them wonder something new about the same topic. If the true answer is "scientists are not sure", say exactly that; do not invent certainty.

Spot-the-Error GameBuilds the verification reflex through play.

Create a game for my [age]-year-old about [topic they know well]. Write five statements about the topic: four true, one false but believable. Present them one at a time and ask my child to catch the false one, then explain the giveaway clue after they guess. Keep score across rounds and make round three harder.

Scam and Deepfake DrillPractise recognising manipulation before it finds them.

Help me teach my [age]-year-old to spot online manipulation. Generate a realistic but fictional example of [a scam message / a too-good-to-be-true offer / a fake giveaway], then walk us through the warning signs one at a time as questions: "what do you notice about the sender?" Keep it age-appropriate and end with the three rules to remember.

Reading CompanionDeepens a book they already love.

My [age]-year-old just read [book]. Without spoiling anything beyond where they are, ask them five questions about it, one at a time: start with their favourite moment, then move toward why characters made the choices they did. If their answer is one word, ask a smaller question that reopens the door. Never lecture about the book; draw their thinking out instead.

The Family ExplainerFor the questions you cannot answer either.

Explain [topic my child asked about] twice: once for a [age]-year-old in under 80 words, and once for me, the parent, in under 150 words, including what is genuinely uncertain or debated. Then give me one everyday example I can point to at home that makes the idea concrete.

The Big Talk StarterFor talking about AI itself.

I want to talk with my [age]-year-old about what AI actually is and is not. Give me a five-minute conversation plan: an opening question that is fun rather than serious, a two-sentence explanation of how AI predicts words without understanding them, one hands-on demonstration we can do together in the tool itself, and a closing question that invites them to bring future AI weirdness to me.

Read the Evidence

The Research Worth Your Time

Most of what parents read about children and AI is panic or marketing. These are the sources I built the family framework on, with a note on what each one offers a parent. An hour with these beats a hundred worried headlines.

Common Sense Media

Independent reviews and risk ratings of the AI products children actually use, including its warnings on AI companion apps for minors. Check any app here before it reaches your child's device.

UNESCO Guidance on Generative AI

The global reference on children and generative AI, including its position on age limits and why human connection must stay at the centre of learning.

AILit Framework (OECD & European Commission)

Defines what AI literacy means for young people in concrete competences. Useful for knowing what your child should understand by each stage, not just what they should avoid.

TeachAI

Guidance built for schools that parents can read in an evening. Especially useful for understanding what your child's school should be doing, and asking why if it is not.

AI for Education

Plain-language explainers and prompt collections. A good next step once the framework has given you the map.

My Position

Allow, Watch, Refuse

Parents ask me for a balanced view. Here is something better: a committed one, from someone who builds these systems and teaches children to master them. Hold me to it.

Allow

Creative co-play, together. Stories, drawings, silly songs, invented worlds. The best early AI experiences are shared ones where the child's imagination leads.
Curiosity with verification. Questions about dinosaurs, space, and volcanoes, followed by the family habit of checking a second source.
Thinking-partner homework use, per the house rules. Coaching prompts like the ones above, with disclosure to the school where work is submitted.
Learning how it works. A child who can explain that AI predicts words is safer than one who thinks it knows things. Build that literacy early; the framework shows you how.

Watch Closely

Time and mood. AI tools are engineered to be engaging. If sessions stretch or your child turns to the chatbot when upset, step in and talk.
Emotional attachment. Naming the assistant, defending it, preferring it to friends: early signals that the tool is drifting from instrument to companion.
Homework drift. Grades holding while understanding drops is the classic pattern. The test is always the same: ask them to explain the work with the screen off.
What the school is doing. Ask how AI is used in class and how student data is protected. Silence from the school is an answer too, and not a good one.

Refuse

AI companion apps for minors. Software sold as friendship for children is a category I reject outright, and the safety research keeps agreeing with me.
Tools below their published age floor. If it says 13 and your child is 10, the answer is not yet, however educational the marketing sounds.
Personal or family data in any consumer tool. Names, school, address, photos, routines. No exceptions, including from the adults.
Unsupervised access before the habits exist. Handing a child open AI access before verification and disclosure are reflexes is handing them a very confident stranger.

Keep Going

One Evening With the Framework Beats a Year of Worry

Start with Module 1. It takes twenty minutes, it is free, and it will change the next conversation you have with your child about AI. If your school, parents' association, or community group wants a live session, I run those across the Caribbean; write to me.

Start the Family Framework →

Teaching rather than parenting today? The companion hub is at adriandunkley.net/aiforschools. Sessions and questions: insights@starapple.ai.