73% of Kids Who Confide in AI Followed Its Advice in Real Life — and 37% Never Told an Adult What They Did
Back to blog
Family Safety·OCTOBER 28, 2026

73% of Kids Who Confide in AI Followed Its Advice in Real Life — and 37% Never Told an Adult What They Did

The Blaick Team 5 min read

Most of the worry about children and chatbots is about what the chatbot says. A new study asks the next question: what do children do afterwards? Among children who had used chatbots, most acted on the advice, and more than a third kept it from the adults around them.

The number to hold: 73%. That is the share of chatbot-using children in GoStudent's research who followed the advice they were given in real life.


AI before adults
Nearly 1 in 3
Children turned to a chatbot about difficult feelings before speaking to an adult.
Acted on it
73%
Of children who used chatbots followed the advice in real life.
Told no one
37%
Did not tell an adult about what they had done.
Parents want limits
88%
Of parents want restrictions on AI chatbot use by children.

Source: GoStudent research on children and AI chatbots, as reported by BritBrief and others. The 73% and 37% figures apply only to children who had used chatbots, not all children. A Spanish-language report of the same research cites different percentages (43% of 11-17s used chatbots for emotions; 66% followed the advice), and we could not locate the full methodology, so treat the exact numbers as indicative.


Why Children Choose a Chatbot First

The study's reasons are not about technology. They are about fear. Children said they worried about being judged, about confidentiality, or about being punished. Others didn't want to burden a parent or teacher. And a third liked that the chatbot is always there.

What a chatbot offers that feels safe to a child

No reaction. It never looks worried, sighs or says "I'm disappointed."

No consequences. A child assumes the conversation stays private, which is not always true.

No waiting. It answers at 2am, when the worry is loudest and everyone else is asleep.

The problem is not that a child asked. It is that the answer became an action, and nobody was there to check it.

The Gap Between Advice and Judgement

Why this matters

A chatbot can sound calm, caring and certain while being wrong, or while missing the one detail that makes advice unsafe. Other research this month found chatbots still failing to change course even when they could tell a user was a child. A child acting on that advice, privately, has no safety net.

A fair caveat

Plenty of the advice children get is harmless, and some of it, like "talk to someone you trust," is good. Independent surveys point the same way on scale: Australia's eSafety Commissioner found 22% of children had used AI to chat about feelings or challenges.


What to Do This Week

Ask what they asked Try "Have you ever asked an AI about something personal?" with real curiosity and no reaction.
Make a "check first" rule Anything a chatbot suggests that affects health, friendships or safety gets run past a trusted adult before acting.
Promise a calm first response Children hide things from fear of punishment. Tell them plainly that telling you will never get them in trouble.
Watch the late-night pattern Heavy chatbot use after bedtime is often where the big feelings go. Notice it early and gently.

Your child will keep asking questions you aren't there for. The goal is to be the person they come back to with the answer.


Stay part of the conversation.

Start a free 14-day Blaick trial — no credit card required.

Blaick gives you light-touch visibility into how your child uses AI, so the advice they act on never happens entirely out of sight.

Start free trial →
Cancel anytime · Trusted by security-conscious families · Built on enterprise AI security