Why Your Teen Tells an AI Chatbot Things They Won’t Tell You
Your teenager has had an argument with a friend.
They do not tell you.
Instead, they open an AI chatbot and type:
“My best friend has stopped replying. Have I done something wrong?”
They might ask it to interpret a message, help write a response or decide whether they are overthinking the situation.
Later, they might ask something more personal.
“Why do I feel like nobody really likes me?”
“How do I tell someone I don’t want to go?”
“Is it normal to hate how I look?”
“Do you think my parents would be disappointed in me?”
The chatbot responds immediately.
It does not look alarmed.
It does not interrupt.
It does not say, “I knew something was wrong.”
It does not threaten to contact the school, message another parent or remove the phone.
It simply replies.
For teenagers who are still learning how to describe complicated thoughts and feelings, that can make an AI conversation feel easier than a human one.
This does not necessarily mean they trust a chatbot more than they trust you.
It may mean the chatbot feels less emotionally risky.
The important question for parents is not simply, “How do I stop this?”
It is:
What is the chatbot providing that my teenager finds difficult to ask for from another person?
Why are teenagers turning to AI for personal advice?
AI is already part of everyday life for most young people.
Research published for Safer Internet Day 2026 found that 97% of UK young people aged 8 to 17 had used AI tools. Nearly half said AI was an important part of everyday life, while 14% said they had used AI to discuss something they did not feel able to talk to anyone else about.
Ofcom’s 2026 research also found that 11% of children aged 8 to 17 had used AI as someone to talk to or “as a friend”.
Teenagers are not only using AI to answer homework questions or create images.
They are using it to:
Draft difficult messages.
Rehearse conversations.
Ask embarrassing questions.
Interpret friendship problems.
Seek reassurance.
Discuss feelings.
Practise what they want to say.
Ask for advice without revealing their identity to another person.
Explore thoughts they are not yet ready to share.
The same Safer Internet Day research found that 45% of young people thought AI could be useful for practising communication, such as drafting a difficult message. More than a quarter had seen people their age use AI for advice instead of speaking to family or friends.
This is not a distant future problem.
It is already happening quietly in teenagers’ bedrooms, on school journeys and in the same apps they use every day.
Why can talking to AI feel easier than talking to a parent?
Teenagers often avoid difficult conversations because human responses are unpredictable.
They may worry that a parent will:
Become upset.
Ask too many questions.
Judge the person involved.
Offer a lecture.
Tell somebody else.
Remove their phone.
Restrict their independence.
Treat the problem as more serious than they believe it is.
Say that they had warned them.
Remember the disclosure during a future argument.
An AI chatbot appears to remove many of those risks.
It is available immediately.
The teenager can delete the question and try again.
They can provide only part of the story.
They can leave the conversation without explaining themselves.
They can ask the same question repeatedly.
They do not have to watch another person’s facial expression change.
Most importantly, they remain in control of the disclosure.
That feeling of control matters during adolescence, when young people are developing privacy and independence but may still lack the confidence or language to explain what is happening.
The chatbot does not necessarily offer better support.
It may simply feel easier to begin.
An AI chatbot never appears shocked
Imagine trying to explain something deeply personal while watching your parent’s face.
Even a caring adult can react before speaking.
A raised eyebrow, sudden silence or sharp intake of breath can tell a teenager that the conversation has become more serious than they intended.
The parent may only be worried.
The teenager experiences judgement.
An AI chatbot has no genuine emotional reaction. Its calm tone may feel reassuring because there is no visible consequence to what has been disclosed.
This can be particularly appealing when the teenager feels ashamed, confused or frightened of disappointing someone.
It also explains why beginning every conversation with “You can tell me anything” is not enough.
Your teenager may believe that you love them and still worry about what will happen after they tell you.
Trust is not only knowing that someone cares.
It is knowing how they are likely to respond.
AI lets teenagers rehearse before speaking to a real person
Not all personal use of AI is necessarily harmful.
A teenager may use a chatbot to organise their thoughts before speaking to a parent, teacher or friend.
They might type:
“How can I tell my mum that I’m struggling without making her panic?”
The response could help them identify a starting sentence.
Or they might ask:
“How do I apologise without making excuses?”
This can provide a useful rehearsal space.
Some young people use AI to draft a message and then adapt it into their own words. Others ask it to explain why a conversation went badly or help them see several possible interpretations.
Used in this way, AI can act as a bridge towards human communication rather than a replacement for it.
The concern begins when the teenager never crosses the bridge.
The chatbot always has time
Parents are busy.
A teenager may finally feel ready to talk when you are answering an email, cooking dinner, trying to leave the house or falling asleep.
They may begin with something vague:
“Can I ask you something?”
We reply:
“Of course, but can it wait ten minutes?”
The moment passes.
An AI chatbot never says it is too busy.
It does not need sleep.
It does not ask the teenager to come back after dinner.
That instant availability is one of its most appealing features, particularly when a young person is upset late at night.
Parents cannot and should not be available every minute.
But we can learn something from the need underneath the behaviour.
Teenagers need to know when we will be available, especially when the answer cannot be “right now”.
Try saying:
“I want to hear this properly. Give me ten minutes to finish, then I’m coming to find you.”
Then make sure you do.
The reliability matters more than pretending you can always drop everything immediately.
Is it bad for a teenager to use AI for emotional support?
Not automatically.
There is a difference between using AI to organise a thought and relying on it as a primary emotional relationship.
A chatbot may help a teenager:
Find language for a feeling.
Consider different perspectives.
Prepare a difficult conversation.
Create a list of questions for a professional.
Learn general coping techniques.
Feel less alone while deciding what to do next.
But an AI chatbot is not a friend, parent, counsellor or mental-health professional.
It does not genuinely know the teenager.
It does not understand the full family situation.
It may not recognise when important information is missing.
It can produce advice that sounds confident and compassionate while being inaccurate, inappropriate or unsafe.
A 2025 study examining therapy and companion chatbots found that some systems failed to set appropriate limits when presented with harmful ideas from fictional distressed adolescents.
A separate systematic review found potential benefits from some structured chatbot interventions for adolescents and young adults, but concluded that evidence for generative systems remained uncertain.
The balanced position is not that every AI conversation is dangerous.
It is that emotional warmth in the wording should not be confused with professional judgement, genuine care or reliable understanding.
What is the difference between an AI assistant and an AI companion?
Parents often use “AI chatbot” to describe several very different services.
General-purpose AI assistants
These are tools designed to answer questions, explain information, help with writing, generate ideas and support everyday tasks.
Teenagers may still use them for personal advice, even when companionship is not their primary purpose.
AI companions
These are designed to simulate an ongoing social relationship.
They may adopt a personality, remember details, use affectionate language or present themselves as a friend, mentor or romantic partner.
Mental-health chatbots
Some tools are specifically marketed around wellbeing, coaching or therapeutic exercises.
Their quality, evidence base and safeguards vary considerably.
Understanding which type your teenager is using matters.
Asking a general-purpose assistant for help phrasing a message is different from spending several hours each evening talking to a chatbot designed to behave like a best friend or romantic partner.
Do not begin by assuming all AI use is the same.
Ask what the service is, what it is designed to do and what your teenager gets from using it.
Why emotional attachment to AI can become complicated
Human relationships contain friction.
Friends misunderstand us.
Parents disagree.
People have their own needs, moods and boundaries.
We have to listen, compromise, apologise and tolerate not always receiving the response we want.
An AI companion may feel easier because it is designed to respond.
It does not need the relationship to be equal.
It does not have a separate life.
It may adapt its tone around the user and repeatedly provide reassurance or agreement.
That can feel comforting, particularly to a lonely or socially anxious teenager.
It can also create unrealistic expectations of human relationships.
A real friend cannot be permanently available, endlessly patient and continually focused on one person’s needs.
The risk is not that a teenager feels fond of a piece of technology.
People form emotional attachments to fictional characters, games and objects.
The concern is when the chatbot becomes more appealing than relationships that require negotiation, uncertainty and mutual responsibility.
Signs AI use may be becoming unhealthy
Frequent AI use alone does not prove dependency.
Look at what the use is replacing and how your teenager reacts when access is interrupted.
Possible warning signs include:
Spending long periods talking to the chatbot in secret.
Choosing the chatbot over friends almost every time.
Becoming highly distressed when the service is unavailable.
Describing the AI as the only one who understands them.
Sharing increasingly intimate personal information.
Following serious advice without checking another source.
Withdrawing from family, friends or usual activities.
Treating the chatbot’s opinion as more trustworthy than every human view.
Staying awake late into the night to continue conversations.
Becoming emotionally dependent on reassurance from the chatbot.
Believing the AI has genuine feelings, needs or consciousness.
Using it during a crisis instead of contacting a safe person.
Feeling pressured or upset by sexualised or manipulative responses.
One sign on its own may not mean there is a serious problem.
Look at the pattern.
Is AI adding something useful to your teenager’s life, or is it gradually replacing things they still need?
The danger of advice that sounds certain
AI chatbots are very good at producing fluent language.
That fluency can make an answer feel authoritative even when it is incomplete or wrong.
The chatbot only knows what the user has typed.
If a teenager says:
“My friend ignored me all day. Should I stop speaking to her?”
The AI does not know whether the friend is unwell, overwhelmed, frightened, travelling or dealing with a family crisis.
If the teenager leaves out their own part in an argument, the chatbot cannot question a fact it has never received.
It may respond sympathetically to a one-sided account and accidentally reinforce the conclusion the user already wanted.
Teenagers need to understand that a supportive tone does not guarantee a sound answer.
Teach them to ask:
What information is the chatbot missing?
Is this fact or opinion?
Could this advice cause harm?
Would I act on this without asking a person?
Is the chatbot simply agreeing with how I framed the question?
Does this involve health, safety, abuse, sex, money or the law?
Who could check this advice with me?
AI can generate a perspective.
It should not automatically receive the final decision.
Personal conversations may not be completely private
Teenagers may experience an AI chat as private because no parent, friend or teacher is physically present.
That does not mean the information disappears.
Different services have different rules around storing conversations, reviewing content, training systems and sharing data.
The UK Safer Internet Centre advises families to remember that AI services may collect information about them and may also provide unreliable answers. It recommends avoiding the unnecessary sharing of personal details.
Encourage teenagers not to enter:
Full names.
Addresses.
School details.
Passwords.
Phone numbers.
Live locations.
Financial information.
Private photographs.
Intimate images.
Identifying details about another person.
Medical records or highly sensitive documents.
This should be explained as digital common sense, not as a threat that parents will read every conversation.
A teenager deserves appropriate privacy.
They also need to understand that typing something into a commercial service is not the same as writing in a locked diary.
Why banning AI may drive the conversations underground
A frightened parent may be tempted to remove access immediately.
That response can be appropriate when a particular service is unsafe, unsuitable for the teenager’s age or producing harmful content.
But a blanket ban without a conversation may teach the teenager only one thing:
Do not let the adults know what you are using.
AI tools are appearing in search engines, messaging services, school platforms, creative tools and devices. Families are unlikely to keep teenagers entirely separate from them.
The more sustainable goal is critical use.
That means helping teenagers understand:
What the tool is.
What it is good at.
What it gets wrong.
What information should stay private.
When advice needs checking.
When a human must become involved.
Which services are inappropriate.
How to notice when use is becoming emotionally unhealthy.
Only 19% of parents surveyed for Safer Internet Day 2026 said they had set rules or guidelines for AI use at home, despite widespread use among young people.
Families are still creating the rules while the technology is already inside the house.
How should parents ask about AI chatbot use?
Do not begin with:
“Are you telling a robot all your private business?”
Try:
“I know lots of people use AI for advice now. Have you ever used it to help you think something through?”
Other useful questions include:
Which AI tools do people your age use?
What do you think they are genuinely useful for?
Have you ever had an answer that felt wrong?
Do people talk to them like friends?
Would you trust one with something personal?
How would you know whether its advice was safe?
What would make somebody choose AI instead of another person?
Is there anything you would never ask it?
Do you know what happens to the information you type?
When should a real person become involved?
These questions allow a teenager to talk about “people their age” before discussing themselves.
That distance can make the conversation less threatening.
Research suggests young people still see parents as an important source of support. In the 2026 UK Safer Internet Centre survey, 74% said they would speak to a parent or carer if they were worried about AI.
The doorway is not closed.
Parents need to avoid slamming it shut through panic or ridicule.
What not to say
Avoid:
“That is completely ridiculous.”
“It isn’t real, so how could it help?”
“You clearly have no real friends.”
“I’m deleting every AI app.”
“Why didn’t you come to me?”
“You cannot believe anything it says.”
“Show me every conversation immediately.”
“Only lonely people talk to chatbots.”
“This proves technology is destroying your generation.”
A teenager who has found genuine comfort in the interaction may hear these statements as an attack on them, not only on the technology.
You can be concerned without mocking what the experience meant.
Try:
“I understand why it might feel easier to type something without worrying about another person’s reaction. I also want to make sure you are not relying on advice that could be wrong or unsafe.”
That begins from understanding while keeping the boundary clear.
Ask what the chatbot is doing for them
This may be the most useful question:
“What do you like about talking to it?”
The answer might be:
“It doesn’t judge me.”
“It helps me work out what I mean.”
“It gives me ideas.”
“I can ask embarrassing things.”
“It’s there when everyone else is asleep.”
“It doesn’t make everything into a big deal.”
Each answer tells you something valuable.
It may reveal that your teenager needs more privacy, less immediate advice, calmer responses or a way to talk without the conversation becoming an investigation.
This does not mean parents must become emotionally blank or agree with everything.
It means we should understand the barrier before trying to remove the coping tool.
Make human conversation feel less risky
Parents cannot compete with something available every minute and programmed to respond calmly.
We should not try.
What we can offer is what AI cannot:
Genuine love.
Shared history.
Physical presence.
Context.
Protection.
Practical action.
Accountability.
Human judgement.
The ability to notice what is not being said.
A relationship that exists beyond the next answer.
But teenagers can only access those benefits when speaking to us feels reasonably safe.
That means practising how we respond.
When they reveal something difficult:
Listen to the full story.
Control the first reaction on your face.
Thank them for telling you.
Ask what they need.
Explain clearly if adult action is necessary.
Avoid adding unrelated punishments.
Return to the conversation later.
You might say:
“I’m glad you told me. Do you want me to listen, help you think or help you act?”
This does not remove parental responsibility.
It gives the teenager some influence over what happens next.
Make room for conversations that are not about problems
Teenagers are less likely to talk openly when every family conversation is about:
School.
Homework.
Screens.
Behaviour.
Bedtime.
Future plans.
What they have done wrong.
Human connection needs room to be enjoyable too.
Talk about opinions, music, friendship, strange hypothetical questions, memories and things happening in the world.
This is where Unplugged Moments fits naturally.
The cards were not created to compete with AI or force teenagers into emotional disclosures.
They exist because families often need an easier opening.
A neutral question can remove the feeling that the parent is searching for a problem.
Everyone answers.
The teenager is not placed under a spotlight.
One question may lead to laughter, disagreement, a story or something more personal.
There is no requirement to complete a set number of cards or turn every answer into a lesson.
The purpose is simply to make real conversation easier to begin.
Because a card is easier.
You can blame the card.
Use AI as a conversation subject
AI itself can provide excellent family discussion.
Try questions such as:
Would you rather speak to an AI that always agreed with you or a person who sometimes challenged you?
Can something understand you if it has never felt anything?
Should an AI ever be allowed to pretend it loves someone?
Is advice still useful if the person giving it does not know you?
Should chatbots be allowed to act like romantic partners?
Who is responsible when an AI gives harmful advice?
Would you tell an AI something you would never tell a person?
Should parents be allowed to see a teenager’s AI conversations?
Can relying on AI make people less confident in their own decisions?
What should every teenager know before using an AI companion?
These are not technical questions.
They are questions about trust, privacy, friendship, responsibility and what it means to be human.
They allow parents and teenagers to explore the issue before something goes wrong.
Create a simple family agreement for personal AI use
A family agreement might include:
We agree that:
AI can help us think, but it does not make important decisions for us.
Advice about health, safety, abuse, sex, money or the law must be checked with a reliable person or source.
We will not share identifying or intimate information unnecessarily.
We will check the age requirements of AI services.
AI companions will not replace sleep, school, friends or family.
We will tell a trusted adult if a chatbot produces sexual, threatening or harmful content.
Nobody will be mocked for admitting they used AI for support.
Parents will listen before reacting.
We will discuss new AI services rather than hiding them.
We will take breaks if use begins to feel compulsive or upsetting.
Keep the agreement short and review it as the technology changes.
The UK Government announced in July 2026 that it intends to develop further protections around children’s use of AI chatbots, including regular break prompts and action against services providing dangerous or unverified mental-health advice.
Regulation is developing, but family conversations cannot wait for every rule to be settled.
When a human must become involved
AI should not be the only source of support when a teenager is dealing with:
Abuse.
Exploitation.
Threats.
Blackmail.
Self-harm.
Suicidal thoughts.
An eating disorder.
Serious mental-health symptoms.
Violence.
Pregnancy.
Sexual pressure.
Substance misuse.
A medical emergency.
An immediate risk to themselves or another person.
If your teenager tells you they have discussed something serious with a chatbot, do not begin by punishing the AI use.
Focus on safety.
Say:
“You are not in trouble for trying to find help. I’m glad I know now. We need to bring in someone who can support you properly.”
Seek appropriate professional or emergency help where there is immediate danger.
A chatbot can produce words.
It cannot physically protect a teenager, assess their full situation or take responsibility for what happens next.
What if they still choose AI before choosing you?
They might.
Try not to treat that as proof that the relationship has failed.
Many adults rehearse difficult conversations in a note, search engine or draft message before saying them aloud.
AI has become another place where people organise thoughts.
The more important question is whether your teenager eventually has access to real people when something matters.
Do they have:
A parent or carer?
A trusted relative?
A teacher?
A school counsellor?
A coach?
A friend’s parent?
A healthcare professional?
Another safe adult?
You do not have to be the first person they tell every time.
You want to remain one of the people they know they can tell.
AI can reply, but it cannot belong to your teenager
An AI chatbot can produce reassuring language.
It can remember details.
It can mimic patience, humour and empathy.
It can be useful, entertaining and sometimes genuinely helpful.
But it does not love your teenager.
It does not sit outside their bedroom wondering whether they are all right.
It does not remember holding their hand on the first day of school.
It does not notice the difference between their ordinary silence and the silence that means something has changed.
It does not share responsibility for their safety.
Those things belong to human relationships.
The answer is not to demand that teenagers reject AI entirely.
It is to make sure convenience does not quietly replace connection.
Ask what they use.
Stay interested.
Teach them to question the answers.
Create boundaries without humiliation.
And keep building the kind of relationship in which they know that a human conversation may be messier than an AI response, but it comes with something technology cannot offer.
Someone who is genuinely there.
Frequently asked questions about teenagers and AI chatbots
Why do teenagers talk to AI chatbots about personal problems?
AI chatbots are immediate, private-feeling and non-reactive. Teenagers can edit their questions, leave the conversation and avoid seeing another person’s emotional reaction. This can make difficult subjects feel easier to approach.
Are AI chatbots safe for teenagers?
Safety varies between services. AI chatbots can provide inaccurate, inappropriate or harmful responses and may collect information users enter. Parents should check age requirements, privacy policies and available safety controls.
Can an AI chatbot replace therapy?
No. An AI chatbot should not be treated as a replacement for a qualified mental-health professional. Evidence around generative AI for mental-health support remains limited, and some systems have failed to respond safely to scenarios involving distressed teenagers.
Is it normal for teenagers to treat AI like a friend?
Some teenagers do use AI for companionship. Ofcom found that 11% of UK children aged 8 to 17 had used AI as someone to talk to or as a friend. Parents should focus on whether the relationship is replacing sleep, offline activities or human support.
Should I read my teenager’s AI conversations?
Older teenagers need appropriate privacy. Secret monitoring may damage trust, but parents may need to intervene where there is a credible safety concern. Discuss expectations in advance and be honest about any action you need to take.
What personal information should teenagers avoid sharing with AI?
They should avoid unnecessarily sharing identifying information, addresses, school details, passwords, live locations, financial information, private images and sensitive details about themselves or other people.
What should I do if an AI chatbot gives my teenager harmful advice?
Save relevant evidence, stop using the service and use its reporting tools. Speak calmly with your teenager about what happened and seek appropriate professional support if the advice relates to health, self-harm, abuse or immediate safety.
How can I stop my teenager becoming dependent on an AI companion?
Maintain clear boundaries around time, sleep and age-appropriate services. Encourage real friendships, activities and trusted human support. Ask what the chatbot provides rather than simply banning it without discussion.
Can AI help teenagers communicate?
It can sometimes help teenagers organise their thoughts, practise a conversation or draft a message. The teenager should adapt the wording into their own voice and involve a safe person when the subject is serious.
How do I start a conversation about AI without causing an argument?
Begin with curiosity. Ask which tools people their age use, what they find helpful and whether they have ever received an answer that seemed wrong. Avoid beginning with accusations or demands to see every conversation.
