Education news

AI Is Flooding Your Child’s Learning. Design Decides If It Teaches.

AI is pouring into your child’s fall this year, and the coverage keeps asking whether screens help or hurt. The learning science says that was never the question. What decides whether a tool teaches is its design, and whether a grown-up stays in the loop.

Laura Lurns Last updated September 3, 2026 · 9 min read

AI Is Flooding Your Child’s Learning. Design Decides If It Teaches.

If your child has come home this fall with a new reading app, a math tutor that talks back, or a homework helper that writes half the answer for them, you are living the story the world’s education ministers are about to spend a week debating. From 8 to 11 September 2026, UNESCO gathers them in Paris for Digital Learning Week under a pointed theme: how learning survives the age of AI. The headlines will sort the tools into good ones and bad ones, screens that help and screens that harm. That sorting is comforting and it is wrong. Whether a tool teaches your child has almost nothing to do with the screen, and almost everything to do with how it was built.

TL;DR

UNESCO’s Digital Learning Week runs 8-11 September 2026 in Paris under the theme ‘Education in the age of AI: Facts, frictions and frontiers,’ with over 1,000 participants and 25 or more education ministers expected to adopt a joint statement.
UNESCO names AI’s opportunities as ‘personalized tutoring and smarter lesson planning’ and its risks as ‘over-reliance, ethical concerns, and declining deep reading skills.’
Learning science finds that tools which make a child retrieve and correct their work step by step (Roediger and Karpicke, 2006; VanLehn, 2011; Brookings, 2025) help more than tools that hand over answers.
Feedback research (Kluger and DeNisi, 1996; Wisniewski, Zierer and Hattie, 2019) finds that a large share of feedback interventions actually lowered performance, so the kind of feedback matters.
In a Harvard trial (Kestin et al., 2025) and a UK math trial run by Google and Eedi (2025), AI tutors built on research-based practices performed well, with the UK trial’s AI supervised by human tutors throughout.

Context the coverage leaves out

What the coverage gets wrong

Most coverage of AI in education sorts itself into two bins: AI will save learning, or AI will wreck it. Both treat the technology itself as the variable, as though ‘a screen’ or ‘an algorithm’ were a single thing that either works or does not. The research supports neither bin. A tool that makes a child retrieve what they are learning and corrects them step by step, with an adult still involved, tends to help; a tool built to maximize time-in-app and badges does not, and both of those are ‘AI on a screen.’ The variable is the design, not the device. The cheating-panic version of the story misses this from the other side: the worry is not that a child touches AI, it is that a tool built to hand over answers lets a child skip the effortful work that learning actually requires. For a parent, that reframes the whole decision. The question is not screen or no screen. It is whether the tool makes your child do the thinking.

AI learning tools are arriving in kids’ lives faster than anyone is testing them, and the good-or-bad framing does not help a parent much. Here are the questions parents are actually asking about what to do at home.

Common questions

Is an AI tutor good or bad for my child?

Neither label fits, because ‘AI tutor’ is not one thing. A tool that makes your child retrieve what they are learning and corrects them step by step, with you or a teacher still involved, tends to help. A tool built to keep your child clicking, counting minutes and handing out badges, does not, even when it looks impressive. Judge the design, not the screen.

How do I tell if a learning app is teaching or keeping my child busy?

Watch what the app makes your child do. If it makes them pull answers out of their own memory, gives specific next-step feedback, and gets harder as they improve, it is doing the work of teaching. If the main thing it produces is time-in-app, streaks and badges, those measure engagement, not whether a skill changed. A child racks up points whether or not anything was learned.

Should an AI tool replace my child’s teacher or tutor?

The evidence points the other way. In the trials where AI tutoring performed best, an adult stayed in the loop, and in one UK study human tutors supervised the AI and edited a meaningful share of its messages. Treat these tools as a supplement to a teaching adult, not a substitute for one. The strongest use frees the grown-up to do the relational work, rather than standing in for them.

Will an AI tool tell me if my child has a learning difference?

No. A learning app or chatbot is not built to diagnose, and you should not read its scores or dashboards as a finding about your child. A screener is a starting point, not a diagnosis. If your child might need formal accommodations such as an IEP or 504 plan, or you suspect a vision, hearing, or medical cause, pursue a professional evaluation too, since that is the only route to those supports.

Share the idea

AI is flooding into your child’s learning this fall. Whether a tool teaches has little to do with the screen and everything to do with its design, and whether an adult stays in the loop.

Post on X Follow

What is happening this fall

From 8 to 11 September 2026, UNESCO gathers education ministers, researchers and technology leaders at its Paris headquarters for Digital Learning Week, under the theme ‘Education in the age of AI: Facts, frictions and frontiers.’ More than 1,000 people are expected on site and thousands more online, and 25 or more education ministers are set to adopt a joint statement on sustaining education as a common good in the age of AI.

UNESCO does not pretend the technology is simple. On its own pages it names the opportunities as ‘personalized tutoring and smarter lesson planning,’ and in the same breath names the risks: ‘over-reliance, ethical concerns, and declining deep reading skills.’ Its guidance for classrooms is to show students how to use AI ‘as a thought partner, not a shortcut.’ The closing keynote, from London School of Economics professor Sonia Livingstone, asks a question a lot of parents are quietly asking too: ‘Is there a right age for AI in education.’

Underneath the summit is the reality in your own home. Reading apps, math tutors, homework helpers and chatbots are arriving in your child’s fall the way tablets did a decade ago, faster than anyone is evaluating them. The question every headline reaches for is whether all this screen-based AI is good or bad for kids. That is the wrong question.

we do not presume that structured AI tutoring will always outperform in-class active learning in all contexts.
Laura Lurns · Learning Success expert

The frame the science supports

Whether a learning tool teaches has almost nothing to do with the screen and almost everything to do with what the tool makes your child do. Cognitive scientists Henry Roediger and Jeffrey Karpicke showed back in 2006 that retrieving what you are learning, struggling to pull it out of your own memory, is what makes it stick, far more than rereading or being shown the answer. Kurt VanLehn’s 2011 review lined up tutoring that responds to a student step by step against software that only checks the final answer, and the step-level kind pulled well ahead. A 2025 Brookings synthesis of several randomized trials put the same split in today’s terms: the AI tutors that helped were built to ‘teach, not tell,’ while students using generative AI with no guidance showed, in Brookings’ words, ‘limited reflection’ on the material.

So the thing to watch for is not the logo on the app. It is a design pattern, and it has a name: engagement mistaken for learning. Time-in-app, lessons completed, badges earned and streaks kept are engagement numbers, and a child racks them up whether or not a single skill has changed. A points system makes a child feel like they are learning. Real mastery makes them feel capable, and those are not the same thing. What actually sustains a child’s willingness to keep engaging is not the next reward, and a tool tuned to maximize clicks is optimizing for the wrong thing.

Two cautions keep this honest. First, feedback is not automatically good: when Avraham Kluger and Angelo DeNisi pooled the feedback research in 1996, a large share of interventions actually lowered performance, a result John Hattie’s team found again across hundreds of studies in 2019. Task-focused, next-step, private feedback helps; a red X, a leaderboard or a bare right-or-wrong often backfires. Second, when a company calls its AI ‘personalized,’ the useful meaning is that it responds to what your child actually got wrong, step by step. It does not mean the discredited idea of matching lessons to a child’s ‘learning style,’ which Pashler and colleagues tested in 2008 and a 2024 review reconfirmed, and which did not hold up.

Key takeaways

Design, not the device: whether a tool teaches depends on what it makes a child do, not the screen.
Engagement is not learning: time-in-app, badges and streaks measure attention, not whether a skill actually changed.
Keep an adult in the loop: the AI tutors that worked best in trials still had a teacher in the loop.

What it means for your child

This is where the news gets practical, because the best real-world evidence we have on AI tutoring says the same thing the older lab work does. In a Harvard study published in 2025, Gregory Kestin and colleagues built an AI physics tutor on a stack of research-based practices, active learning, scaffolding and targeted feedback among them, and it outpaced a well-taught active-learning class. That is a genuinely encouraging result, and it is also one course over two weeks, which the authors are careful to say does not prove the pattern holds everywhere. In a 2025 UK trial run by Google and the Eedi math platform, an AI tutor also performed well, but with a crucial condition: human tutors supervised it throughout and edited a meaningful share of its messages. The AI that worked was the AI with an adult still in the loop.

The failure mode is as instructive as the success. When university students in a 2026 study used an unrestricted chatbot to do an assignment, they scored worse on a surprise test weeks later than students who studied the ordinary way. That study was on adults, not children, and it was open-ended chatbot use rather than a designed tutor, so read it narrowly. What it illustrates is the mechanism underneath all of this: when a tool does the effortful thinking for you, you pay for it later in what you failed to learn. That is true of a screen and equally true of a worksheet someone else fills in.

So the parent move is not to ban the screen and not to trust it either. It is to ask one question of any tool your child uses: is this making my child do the thinking, or doing it for them? A tool that makes a child retrieve, corrects them step by step, and leaves you or a teacher in the loop is worth keeping. A tool that hands over answers and counts the minutes is not, however impressive the technology looks. If screen habits are already a worry in your house, the same standard sorts the digital tools worth your child’s time from the ones burning it.

showing students how to use AI as a thought partner, not a shortcut.
Laura Lurns · Learning Success expert

You do not need a degree in learning science to make a good call about a screen in your child’s hands. You need one question: is this tool making my child do the thinking, or doing it for them? Children are capable of the effortful work that actually builds a skill, and a tool designed to keep them clicking quietly takes that work away from them. The villain here is not AI, and it is not the screen. It is an old design trick wearing new technology: engagement mistaken for learning, where minutes logged and badges earned stand in for whether anything was learned at all. That is the pattern Learning Success was built to refuse. Our All-Access membership opens an assessment that asks about the processing systems your child’s learning runs on, and a roadmap that names which skills to build first and in what order. It is a starting point for helping your child, not a diagnosis.

See what All Access gives your child

Is your child struggling in school?

Get your free personalized learning roadmap

You describe what you see at home. We turn it into a plan you start this week.

Answer 5 short questionnaires about what you already notice, 30–45 minutes at your own kitchen table
Your child sits no test and gets no score: nothing to schedule, nothing for them to dread
You do the answering, the AI does the writing, and a person reviews it before it reaches you
Access all 40+ courses instantly: reading, math, focus, processing and more, with new ones added regularly

Why we use AI, plainly: it writes from a knowledge base our team maintains and audits. We work through it line by line and pull anything the evidence stops supporting. The roadmap you get on Tuesday reflects what we corrected on Monday, and a human still reads it before you do.

Your school district must evaluate your child free of charge if you ask in writing, whatever your income and whatever the outcome (US, 34 CFR 300.111 and 300.301(b)). That route takes time and answers a different question than you do. This one starts today, from what you already know.

Get my free assessment & 7-day trial Cancel anytime during your trial.

Your answers stay yours. We do not sell your personal information, and we do not hand identifiable assessment data to outside AI companies to train their models.

A screener is a starting point, not a diagnosis. If your child might need formal accommodations (an IEP or 504 plan), or you suspect a vision, hearing or medical cause, pursue a professional evaluation too. That is the only route to those supports.

References

UNESCO — Education in the age of AI: ministerial statement to be adopted at UNESCO’s Digital Learning Week
UNESCO — Artificial intelligence and education
Kestin et al., Scientific Reports (2025) — AI tutoring outperforms in-class active learning: an RCT
LearnLM Team, Google & Eedi (2025) — AI tutoring can safely and effectively support students: an exploratory RCT in UK classrooms
Brookings (2025) — What the research shows about generative AI in tutoring
ScienceAlert — Coverage of Barcaui (2026): a randomized trial on ChatGPT use and knowledge retention

Laura Lurns · Learning Success expert Writes about the learning brain for parents who want plain answers. Every article is grounded in current neuroscience and classroom practice.