Start learning

For parents and educators

AI for teenagers: what parents should know

Your teenager is already using AI, whatever the household rules say. The useful question is whether they can tell when it's wrong — and that's a skill you can teach, not a personality trait.

Worth worrying about

Thinking, outsourced

Not homework done by a machine — that's the visible version, and schools are dealing with it. The subtler problem is a student who never experiences not knowing something, because the answer is always thirty seconds away.

The struggle isn't a side effect of learning, it's the mechanism: effortful recall is what tells the brain something is worth keeping. A teenager who has never sat with confusion hasn't built any tolerance for it, and every hard subject eventually asks for that tolerance.

Plausible wrongness

Teenagers were taught to spot the obviously unreliable: the badly made website, the anonymous forum post, the page covered in ads. AI output has none of those tells. It's articulate, confident and well organized whether or not it's correct.

A table of the warning signs teenagers are taught to look for — spelling mistakes, an ad-covered page, no named author, copied text. Every one of them shows up on a sketchy website and none of them show up on an AI answer. Sketchy website AI answer Spelling mistakes Ads all over the page No named author Copied from nowhere
A check means the warning sign is there to be spotted. Every heuristic they were taught fires on the left and none of them fire on the right.

Calibration, in both directions

Some teenagers trust it completely. Others, having been warned, refuse to touch it. Both are blanket rules adopted to avoid making a judgment. What you want is a young person who can say "I'd trust it for this and check it for that", and say why.

Worth worrying about less

  • "They won't learn to write." Writing is thinking, so the argument is sound — but the response is to make the thinking visible: drafts, revisions, explaining choices out loud. We had the same argument about calculators.
  • "They'll fall behind on the newest tool." The tools turn over every few months; the ideas underneath don't. Someone who understands how these systems work picks up a new interface in an afternoon.
  • "They need to specialize now." They don't. Breadth plus real understanding beats early narrow expertise in a field where the specifics turn over this fast.

Questions worth asking at home

These work better than "did you use AI for that?", which only teaches them to be quieter.

  • "How would you know if it had got that wrong?"
  • "What did it get wrong recently?" — a teenager who can't name an example isn't checking.
  • "Explain that bit to me like I'm clueless." No jargon allowed. It's the fastest test of whether the understanding is real, and it puts you in the pleasant position of being taught.
  • "What would you have done before you had this?"
  • "Where would you not use it?" — the answer to this one tells you how they're doing.

What Toorova does

  • 13 and up. The AI features are never available to an under-13 account, by anyone including a parent, and that's enforced on our servers rather than in the app.
  • For 13 to 17, you approve, from your own account. The permission is yours to give and to withdraw, not a box ticked on their device, and it stops applying if the guardian link is removed.
  • The course runs on getting things wrong first. A question comes before the teaching so the student finds out what they don't know, and a wrong answer is told why it was tempting.
  • Passing is separate from finishing. Certificates require passing every lesson they cover, so tapping through earns nothing.

Why you can't see their wrong answers

Two columns. A linked parent sees streak, XP and level, lessons finished, lessons passed and the course they are on. A linked parent never sees their answers, their wrong guesses, journal notes or tutor conversations. What you see What you never see Streak, XP, level Lessons finished Lessons passed Course they're on Their answers Their wrong guesses Journal notes Tutor conversations enforced in the database
The line between the two columns is enforced where the data lives, not hidden on the screen, and it isn't configurable.

The method depends on a learner being willing to guess wrong. A teenager who believes every wrong answer is reported home stops guessing and starts playing safe, and at that point the dashboard stops telling you anything true. You get honest signal about whether they're learning because they aren't being watched at the level of individual mistakes.

What's stored and who can see it is in the Privacy Policy; linking and unlinking are covered in the help center.

Next