Before Your Child Uses AI: What Should Never Be Shared
A child can ask an AI tool for help with homework, a story idea, a Bible verse explanation, a recipe, or a hard friendship question. In seconds, it can answer with confidence. That speed can feel helpful, even magical.
But children also need to learn this early: AI does not know them, love them, protect them, or carry responsibility for their well-being.
That is the work of trusted humans.
This guide from the Sister Georgie Healing Library is part of our October series, Safe, Seen & Strong: Raising Children in a Digital World. September centred parenting awareness, conflict, and healing homes. October moves into the daily environment many children now live in, with phones, social media, online conflict, and artificial intelligence.
The World Health Organization reports that one in seven children and young people aged 10 to 19 experiences a mental disorder globally. WHO’s review of digital technology and youth also reminds us that technology’s effects are mixed. The impact depends on the child, the context, the support around them, and any existing vulnerability.
So this is not a panic piece.
AI can support learning. It can explain a maths problem in a new way, help a child practise writing, translate a sentence, or spark curiosity. But families need clear boundaries. Children should know what never belongs in a prompt, what answers need checking, and when only a human being will do.

AI is a tool, not a trusted adult
AI tools can sound kind. They can write in a friendly tone. Some even respond as if they understand feelings. That can make them feel safe to children.
But AI does not understand the way a caring adult understands.
It predicts and generates responses based on patterns in data. It may not know whether a child is scared, confused, being bullied, or giving away too much personal information. It may not know what is culturally sensitive, age-appropriate, or emotionally safe for that child in that moment.
UNICEF’s child-centred guidance on AI places children’s best interests at the centre. It highlights safety, privacy, fairness, transparency, accountability, healthy development, and protection from harm. In family language, that means:
Children should not be treated as testing grounds.
Tools should protect children’s data and dignity.
Systems should be fair and not discriminate.
Children should be able to understand when AI is being used.
Adults, companies, schools, and governments must remain accountable.
A child’s growth, wellbeing, and safety must come before convenience.
At home, that begins with one plain message:
AI is a tool. Wisdom still needs a human.
That sentence is worth repeating until it becomes normal.
What children should never share with AI
Children may not understand how much information a few small details can reveal. A prompt that seems harmless can include enough to identify them, locate them, embarrass them, or expose a family problem.
Teach children that AI tools are not private diaries, counsellors, confession boxes, or best friends.
Here are the main things children should not casually share.
Full names and identifying details
A child should not enter their full name, school name, home address, phone number, email address, family members’ names, or usernames.
Even if the tool seems harmless, children should learn the habit of keeping identifying details out of prompts.
Instead of typing:
“My name is [full name], I go to [school], and I need help writing about my class trip.”
They can write:
“I am a student writing about a class trip. Help me plan three paragraphs.”
The help can still happen without revealing who they are.
Photos, videos, and voice recordings
Children may want to upload a selfie, a school photo, a video from home, or a voice note. They may not realise that images can reveal uniforms, locations, faces, bedroom details, routines, or other people who did not give permission.
A simple family rule can help:
Do not upload your face, your school uniform, your home, or anyone else’s face unless a trusted adult checks first.
This includes younger siblings, cousins, classmates, and friends.
Private family information
Children should not share family conflict, illness, financial stress, immigration matters, custody issues, addresses, passwords, or private conversations with AI.
They may ask because they want comfort. That is understandable. But AI can give advice that is incomplete, insensitive, or unsafe.
A safer prompt might be:
“I feel upset after an argument at home. Can you suggest calm breathing exercises until I speak to a trusted adult?”
That keeps private details out and still points the child toward support.
Secrets, shame, and deeply personal feelings
AI may feel easier to talk to than a person, especially for a child who feels embarrassed. But if a child is experiencing fear, abuse, bullying, self-harm thoughts, sexual pressure, exploitation, or deep distress, they need a safe human.
A child can be taught this rule:
If the problem feels too heavy to say out loud, it is probably time to find a safe adult, not a chatbot.
AI can suggest general coping steps. It cannot replace a parent, guardian, teacher, counsellor, doctor, pastor, youth worker, or emergency support.
Passwords and account information
This one should be clear and repeated often.
Children should never type passwords, PINs, verification codes, bank details, app logins, or private account information into AI tools. They should also avoid asking AI to “check” a password, remember a code, or write a message that includes login details.
No homework help is worth losing access to an account or exposing a family device.

The PAUSE toolkit before every prompt
Children need simple tools they can remember. The PAUSE Before You Prompt toolkit gives families a quick check before using AI.
PAUSE stands for:
Letter | Question | What it teaches |
P | Private? | Am I sharing something that belongs to me, my family, my school, or someone else? |
A | Accurate? | Could this answer be wrong, made up, old, or missing context? |
U | Understandable? | Do I actually understand the answer, or does it just sound smart? |
S | Safe? | Could this advice hurt me, embarrass someone, or put anyone at risk? |
E | Explainable? | Can I explain where this answer came from and why I trust it? |
This toolkit works best when practised out loud.
A parent might say:
“Before we ask, let’s PAUSE. Are we about to share anything private?”
A child might answer:
“No names, no school, no address, no faces.”
That short exchange builds a habit. Over time, children learn that digital safety is not only about rules. It is about judgment.
AI can be wrong even when it sounds sure
One of the hardest things for children to understand is that AI can be confident and incorrect at the same time.
It may:
Give a wrong fact.
Make up a source.
Misread a question.
Leave out important context.
Create an image that looks real but is not.
Give advice that does not fit the child’s age, culture, health, or situation.
This matters because children often trust polished language. If an answer sounds neat, they may assume it is true.
Families can normalise checking.
For schoolwork, children can compare AI answers with textbooks, teacher notes, trusted educational websites, library books, or a knowledgeable adult. For health, safety, legal, family, or emotional issues, AI should not be the final voice.
Try this family phrase:
AI can help us start. It does not get the last word.
That keeps the door open to learning while protecting children from false certainty.
It also protects integrity. If a child uses AI to write an entire assignment and hands it in as their own work, they miss the learning. A better use is to ask AI to explain a concept, give practice questions, suggest an outline, or help them review their own draft.
The goal is not to ban help. The goal is to keep the child thinking.

Children need to know when a human is needed
AI can answer many questions, but some situations need a living, caring person.
Teach children to stop using AI and seek help from a trusted adult when:
Someone online asks them to keep a secret.
They feel scared, trapped, ashamed, or threatened.
They are being bullied or pressured.
They see or receive sexual content.
They are asked to send photos, videos, money, passwords, or personal details.
They are thinking about hurting themselves or someone else.
They need medical, legal, or serious safety advice.
They do not understand what the AI is asking them to share.
The adult does not have to be perfect. The child simply needs a safe pathway. Families can name three trusted adults together, such as a parent, grandparent, auntie, uncle, teacher, guidance counsellor, pastor, coach, or family friend.
Children should also know that asking for help will not lead to instant punishment. If children believe they will be shamed or lose all device access, they may hide the problem. A calm first response makes future honesty more likely.
Try saying:
“Thank you for telling me. We will work through this together.”
After that, adults can set limits, report harm, save evidence, contact the school, or seek professional support if needed.
This article is for general information and family education. It is not a substitute for professional mental health, medical, legal, or emergency advice.
Five parent-child questions for evaluating an AI tool
Before a child uses a new AI tool, sit together and ask these five questions. They are based on the values UNICEF highlights in child-centred AI guidance, including safety, privacy, fairness, transparency, accountability, development, and children’s best interests.
1. What does this tool ask us to share?
Look at sign-up forms, prompts, uploads, and permissions. Does it ask for a name, age, email, camera access, microphone access, location, photos, or contacts?
If the tool asks for more than it needs, slow down.
2. Can we tell when AI is being used?
Children should know when they are talking to AI, reading AI-generated text, or seeing an AI-made image. A tool should not trick the child into thinking it is a human friend.
Transparency matters because children deserve to understand what they are interacting with.
3. Who checks whether the tool is safe for children?
Ask whether the app or website explains its child safety measures in plain language. Does it offer reporting tools? Does it have age guidance? Does it give parents or guardians clear controls?
If no one appears accountable, that is a warning sign.
4. Could this tool treat people unfairly?
AI can reflect bias from the data used to build it. It may give unfair, stereotyped, or harmful responses about race, gender, disability, language, nationality, religion, body size, or family background.
Ask the child:
“Does this answer respect people? Does it make unfair assumptions?”
That builds moral awareness, not just tech skill.
5. Does this tool help the child grow?
A good tool should support learning, creativity, curiosity, and confidence. It should not encourage cheating, isolation, harmful comparison, secrecy, or dependence.
Ask:
“After using this, do I understand more, or did it do all the thinking for me?”
That question helps children see the difference between support and replacement.

A simple family agreement for AI use
A written agreement can make expectations clear without turning every AI question into an argument. Keep it short enough for a child to remember.
Here is a starter version families can adapt:
I will not share my full name, address, school, phone number, passwords, photos, or private family information with AI.
I will ask an adult before uploading images, videos, voice notes, or documents.
I will check important answers with a trusted person or reliable source.
I will not use AI to bully, embarrass, trick, or copy someone else’s work.
I will tell a trusted adult if something online makes me feel scared, pressured, confused, or unsafe.
I will remember that AI is a tool, and wisdom still needs a human.
For younger children, use simpler wording:
“Do not tell AI who you are, where you are, or things that can hurt your heart. Ask a grown-up.”
For older children and teens, invite discussion. They may already be using AI at school, on phones, or through search tools without naming it as AI. Ask what they have seen. Ask what feels useful. Ask what feels strange.
Respectful conversation gives parents more influence than fear-based control.
Keep the door open
The strongest digital safety lesson is not “AI is bad.” That message will not hold up in a world where AI appears in search engines, school tools, apps, games, study platforms, and creative software.
A stronger message is this:
AI can help, but it cannot be trusted with everything.
Children need boundaries that protect their privacy. They need skills to question answers. They need adults who will listen before reacting. They need tools that are designed with their best interests in mind. They also need space to learn, create, and ask good questions.
Start with one family conversation this week. Choose one AI tool your child has used or heard about. Sit together, try the PAUSE questions, and agree on what should never be shared.
Safe children are not children who never touch technology. Safe children are guided, seen, and strong enough to know when to stop, check, and call in a human.







Comments