Your Kid Is Using AI for School Every Day. Here Is the One Skill Almost No One Has Taught Them.

There is a number buried in Microsoft’s new AI in Education Report that has stayed with me since I read it. Ninety-two percent of students now use AI for school.^[1] That was not the surprising part. The number next to it was. Seventy-seven percent of those same students say they have never received formal training on how to use it.^[1] That means the tool your kid uses every day for homework is one almost no one has actually taught them to use. That gap does not close on its own.

The gap between how many kids use AI for school and how many have learned to use it well has never been wider.

The Microsoft survey covers students, educators, and school leaders across multiple countries, published June 24, 2026. Its headline finding: 92% of students have already used AI for school purposes.^[1] That means AI use for schoolwork is not a trend that is still arriving. It is the default for the overwhelming majority of kids. Among those students, 65% say their AI use for school increased over the past year.^[1] This is happening faster than most schools, most curricula, and most parents expected.

A separate UNICEF analysis, drawing on data from children across 10 countries, found that kids are adopting AI at rates more than three times faster than adults.^[2] Kids are not waiting for permission or guidance before picking up these tools. They are using them the way every generation has used new technology: through trial, peer modeling, and whatever habits happen to work first. The difference with AI is that the habits forming now are habits about how to think, not just how to do.

Here is the number I keep coming back to. Seventy-seven percent of students say they have never received formal AI training.^[1] Not a class, not a session, not a structured conversation about how to use the tool well. More than three in four kids using AI every day for school have been left to figure it out on their own. Schools are genuinely trying to catch up. But the habits are forming right now, and the instruction is not yet there.

Habits formed without instruction tend to form around whatever happened to work first, and that is not the same as learning to use a tool well.

My 13-year-old runs her school life with the focus of someone in training. She is also, I realized recently, using AI in a way that looks nothing like what I would want to teach her. Not because she is doing something wrong. Because nobody has taught her what good AI use actually looks like, and she has not needed to know until now. What she has developed is a set of habits based on efficiency and ease. Those habits are consistent and practiced. They are also entirely self-taught, and that matters.

The Microsoft report found that 41% of students cite academic integrity as one of their top concerns about AI.^[1] What that number tells me is that kids already sense something is off. They know there is a line somewhere. They just have not been given a map of where it is. Without that map, they are drawing it themselves, under deadline pressure, surrounded by peers doing the same thing, with no shared standard to work from.

The deeper issue is not dishonesty. It is dependency. A student who consistently turns to AI before attempting a problem, forming an opinion, or drafting a first thought is not learning to think through hard things. They are learning to avoid the discomfort of not immediately knowing. That avoidance, practiced enough times, becomes default behavior. It does not feel like a bad habit from the inside. It just feels efficient.

Intentional Use means deciding, before opening a chatbot, what you are trying to learn and what role AI will play in getting you there.

Intentional Use is not about restricting AI. That conversation is already losing, and most parents know it. It is the difference between a kid who reaches for a chatbot because an answer feels hard and a kid who reaches for one with a clear plan: I am going to ask it to explain this concept, compare its answer to what I already know, and then write my own version in my own words. Same tool. Very different habit. The outcome of the second approach is not just a completed assignment. It is a kid who learned something.

Educators who are introducing AI to elementary students have landed on a similar insight. One teacher in Massachusetts, quoted in Education Week this week, put it plainly: “Teach them good habits now, so middle school and high school teachers don’t have to teach them to break their habits.”^[3] The window for building intentional habits is not infinite. The longer a kid uses a tool one way, the more automatic that way becomes. Automatic is not the goal here.

Intentional Use also includes what you do after AI responds. Not accepting the output because it arrived in complete, confident-sounding sentences. Reading it with a question in mind. Noticing when something feels off. Asking a follow-up before you move on. UNICEF found that a third of children in their survey worry about AI being used to spread misinformation and scams.^[2] That instinct is healthy. Intentional Use is what turns that worry into a habit of actually checking.

The clearest sign is when your kid treats AI output as a starting point, not a stopping point.

Watch for this: your child finishes something AI-assisted and then talks about what they changed. Not just “I used ChatGPT” but “I asked it for the main points, then rewrote the explanation because the way it said it was confusing.” Or they catch an error and mention it unprompted. Or they notice the AI gave a different answer than what their teacher said and they are not sure which one to trust. Any of these is the skill in motion. They are evaluating, not just consuming.

My middle daughter did something last month that I found myself thinking about for days. She had used AI to help with a social studies assignment and noticed the answer was missing something her teacher had mentioned in class. She came to me not to complain about the AI but to ask which source she should trust. She did not know the answer. But she had caught the gap between what the AI told her and what she already knew. That small moment, noticing the inconsistency and taking it seriously, is what Intentional Use looks like when it is still forming. The instinct to ask “wait, is that right?” is worth everything. Encourage it every time you see it.

You don’t need to understand AI better than your kid does. You need to ask the questions that make them slow down and reflect on what they did with it.

Ask “what did you do with what it gave you?” after any AI-assisted task. Not “did you use AI?” That question invites defensiveness. The better question treats AI use as normal and asks about the follow-through. Did they check anything? Did they rewrite anything? Did something surprise them or seem wrong? The question trains the habit of looking critically at AI output rather than just accepting it.

Do one AI task together as a beginner this week. Pick something you genuinely do not know the answer to, a recipe adjustment, a historical question, a local fact, and use a chatbot together. When it gives you an answer, narrate your thinking out loud: “I am going to check that part.” “That seems right but I want to confirm it.” “I would say that differently.” Showing what Intentional Use looks like is more effective than explaining it, especially with older kids who have already formed their own habits.

Ask their school for its AI policy in writing. The Microsoft report found that 42% of educators say they have no clear classroom guidelines to give students.^[1] If your school does not have a policy, that is information you need. Not so you can deliver a long lecture at dinner. So you can have a short, honest conversation about what you both think good AI use actually looks like. That conversation, even without a clear conclusion, starts building the skill.

Eighty-seven percent of students in the Microsoft survey agree that knowing how to use AI well matters for their future.^[1] They are not waiting to be convinced. They already believe it. What they are waiting for is someone to show them what good actually looks like. You do not need credentials for that. You need to be curious alongside them, willing to say “I don’t know” when you don’t, and willing to model the checking and questioning you want them to do on their own.

I have been thinking about this through the lens of how I learned to use a library as a kid. Somebody taught me that the first source you find is not necessarily the right one, that you check the publication date, and that you look for who is making the claim before you repeat it. AI is a fundamentally different tool. The underlying skill is the same.

The training gap in the Microsoft data is not a reason to panic. It is a map of where parents are actually needed right now. Schools will catch up. The research on how to teach AI well is being written in real time by educators who are figuring it out in their classrooms, just like kids are figuring it out at their desks. In that gap, the guidance that matters most is not expert instruction. It is a parent who asks the right questions and is honest about not having all the answers.