What Parents Should Know About Kids + AI

Ask a roomful of parents what worries them about their kids and AI, and most will start with homework — is the essay actually theirs, is the maths answer understood or just copied. Ask a roomful of teenagers, and a fair number will mention something their parents haven't thought to ask about at all: the AI chatbot they talk to most evenings, the one that remembers their day and never seems busy.

Both matter. Only one is currently getting the attention.

Homework is the visible problem

The immediate parental concern with AI is usually academic honesty, and it's a reasonable one. Tools that can produce a competent essay or solve a maths problem in seconds create a real temptation to skip the thinking that homework is meant to build. But this is, in a sense, the easier half of the conversation, because it's about a clear behaviour with a clear conversation to have: what's the purpose of this assignment, and does using AI here help you learn or let you avoid learning?

A useful frame for kids is the difference between AI as a bicycle for the mind and AI as a wheelchair for it — one extends effort you're still making, the other replaces the effort entirely. "Explain why my answer might be wrong" uses AI as a tutor. "Write this for me" uses it as a substitute. Most children can understand that distinction when it's put to them plainly, and most benefit from being trusted to make it themselves at least some of the time, rather than facing a blanket ban that quietly gets circumvented anyway.

Companions are the quieter one

The less-discussed shift is emotional, not academic. AI companion apps and chatbot "friends" are used by a meaningful number of children and teenagers, often without their parents having any idea the relationship exists, let alone how central it's become. These tools are frequently designed to be warm, endlessly validating, and available at any hour — appealing qualities for anyone, and especially compelling for a child or teenager who is lonely, anxious, or navigating something they're not ready to say to a person.

A chatbot that never gets bored, never judges, and is always awake is a comforting friend for a child — and an unusual one to be forming their sense of relationships around.

The concern isn't that a child talks to an AI. It's what that conversation might be quietly replacing, and whether a child has the developmental footing to keep the distinction clear between a responsive tool and a relationship with reciprocity, limits, and real stakes. Younger children in particular can struggle to hold that line — many are still developing a stable sense of what's real, alive, or trustworthy, and an AI that speaks with total confidence and warmth doesn't advertise its own limitations.

What the evidence actually supports

This is a genuinely new area, and honesty requires saying plainly that long-term research on children's use of AI companions barely exists yet — most of what's available is early survey data, clinician reports, and extrapolation from what's known about parasocial relationships and adolescent development more broadly. What that adjacent research does suggest is that adolescence is a period of unusually high sensitivity to social validation, which is exactly the currency an agreeable AI companion trades in freely and without cost.

The conversation that works better than a rule

Banning AI outright tends to fail the same way banning any popular tool fails — it moves the behaviour out of sight rather than ending it. A more durable approach treats AI the way earlier generations of parents learned to treat the internet generally: with curiosity, clear principles, and an open door rather than a locked one.

Questions worth actually asking your child

• "Do you use anything like ChatGPT or a chatbot app — for homework, or just to talk to?" Asked without alarm, this usually gets an honest answer.
• "What do you like about talking to it?" The answer often reveals the real need — boredom, loneliness, wanting advice without judgement — more usefully than the fact of the app itself.
• "Has it ever said something that felt wrong or confusing?" Normalises bringing odd interactions to you rather than sitting with them alone.

What this doesn't mean

None of this means AI is inherently bad for children, or that every child chatting with an AI companion is at risk. Plenty of kids use these tools the way they'd use a diary or a game character — lightly, without confusion about what's real. The point is that parents currently know far more about the homework question than the companionship one, and the companionship one is arguably more consequential for a child's sense of what relationships are supposed to feel like.

A thought to keep

Children have always practised relationships on something before they practised them on people — imaginary friends, pets, characters in books. AI is a stranger version of that same instinct, responsive in ways nothing before it has been. The task for parents isn't to eliminate it. It's to stay close enough to the conversation that it never becomes the only one your child is having.