Schools Scale AI Fast. 77 Percent of Students Get No Training to Use It.

Last updated: July 4, 2026

Your child’s school added an AI tool this year. That part is almost certainly true — 92% of students and education leaders now report using AI for school purposes, and school leaders are moving fast to add more. Here is what most schools did not do before the rollout: train anyone. A new report from Microsoft, conducted by PSB Insights across 3,345 students, educators, and leaders in six countries, found that 58% of school leaders are already scaling AI — and 77% of students have received no formal training on how to use it. Fifty-three percent of educators are in the same position. The tool arrived. The preparation did not.

TL;DR

Microsoft’s 2026 AI in Education Report (3,345 respondents, 6 countries) found 92% of students have used AI for school purposes — a 26% rise in one year.
77% of students and 53% of educators have not received formal AI training, despite 58% of school leaders actively scaling AI use.
Only 31% of U.S. public schools have a written AI use policy; 30% have no plans to create one (U.S. Dept. of Education, December 2024).
In high-poverty schools, 39% of teachers have AI training vs. 67% in low-poverty districts; RAND projects the gap will persist.
66% of educators want monthly or quarterly AI training from their institutions — demand that institutions are not meeting.

Microsoft’s 2026 global survey found that AI is in nearly every school — and 77% of students never received training to use it. Here is what that gap means for the kids who were already falling behind.

Common questions

Is AI in school automatically good for kids who struggle with reading or learning?

Not automatically. AI tools vary widely, and the evidence base for most edtech tools is thin. The key variable is not the tool — it is whether teachers have been trained to use it effectively and whether the tool supplements skilled instruction or replaces it. A well-supported program at a prepared school has potential. An AI tool in an untrained teacher’s hands, without a policy framework, adds a variable rather than removing one.

How do I find out if my child’s school has an AI policy?

Ask the principal or curriculum director directly: “Does the school have a written policy governing students’ AI use, and have teachers received formal AI training?” As of December 2024, only 31% of U.S. public schools had a written policy. The absence of one does not mean AI is absent from classrooms — it often means it arrived without a plan. That is information worth having before your child uses these tools daily.

My child has an IEP. Should AI tools be part of the plan?

It depends on the tool, the goal, and whether the teacher using it with your child has been trained to do so. AI can be written into an IEP as an accommodation — text-to-speech, a writing aid — but the question special education research asks about every accommodation applies: is this building the underlying skill, or replacing the expectation that it gets built? 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 — that is the route to those supports.

What is the difference between using AI and understanding AI?

A significant one. Using AI means knowing how to prompt a chatbot or run an AI-assisted reading program. Understanding AI means knowing how these systems produce outputs — what their error patterns look like, when to trust them, and how to push back when they get something wrong. For a child who already carries fragile confidence around reading or math, the difference between accepting an AI answer and evaluating one is the difference between a tool that helps and one that quietly confirms the wrong thing. Building that habit is something parents can start at home right now, well before the school catches up.

Schools are scaling AI fast. 77% of students never got training to use it. Teachers in high-poverty schools are half as likely to be trained as those in low-poverty schools. We have run this exact play before. Tweet Now Follow Learning Success

What the Microsoft Report Found

Released June 24, 2026, the Microsoft AI in Education Report surveyed students, educators, and leaders across the United States, United Kingdom, Australia, Brazil, Japan, and Saudi Arabia. AI use rose 26% among students and 21% among educators in a single school year — a pace that makes even the smartphone’s classroom arrival look measured. But inside that surge sits a structural failure. Despite near-universal adoption, 77% of students and 53% of educators have not received formal AI training. Only 31% of U.S. public schools have a written policy governing how students use AI, according to U.S. Department of Education data. Thirty percent of schools report having no plans to create one.

This is not a supply problem. The demand for training is clear and documented: 66% of educators say they want structured AI guidance from their institutions monthly or quarterly. Fifty-two percent of students say the same. The institutions have not delivered it. A concurrent RAND Corporation analysis sharpens the picture. In high-poverty schools — where students with learning differences are disproportionately concentrated — only 39% of teachers have received any AI training at all. In low-poverty districts, that number is 67%. RAND projects the gap will persist as training programs expand slowly.

What the coverage gets wrong

Most coverage frames AI adoption statistics as a success story — “92% of students use AI” becomes the headline, and the 77% training gap is buried or called a “next step.” Prior education technology adoption cycles show that adoption pace is not the same as educational impact, and that students who benefit least from a tool adopted without training are those who most needed skilled instruction in the first place. The RAND data make the equity dimension concrete: the children in the highest-need schools face the largest training gap. The real story is not that AI arrived in classrooms. It is that in the schools where it matters most, it arrived before the readiness did.

We Have Watched This Movie Before

Education has a documented pattern here, and naming it plainly is more useful than calling it a growing pain. Researchers proved “learning styles” instruction doesn’t help kids learn back in 2008, and again in 2024. A review of educators across 18 countries in 2020 found nearly 9 in 10 still teach to them. Cognitive scientists knew for decades that teaching children to guess words from pictures and context trained them to read like struggling readers, not strong ones. It took a national investigative podcast in 2019 to start changing laws, and most of those laws are only a year or two old. The science was not the problem. The system that adopts new ideas faster than it trains the people holding them was.

AI is now moving through that same system at a pace that makes those earlier adoption waves look cautious. The gap is not a failure of individual educators — 66% of them are actively requesting better support from their institutions. It is a structural choice: treating access to a tool as equivalent to readiness to use it. And the children who pay for that choice are the ones who were already being underserved. A teacher in a high-poverty school, handed a new AI tool without training, is not better positioned to support a child who already struggles with reading, attention, or processing than she was before the tool arrived. She is busier. The compounding disadvantage arrives, as it always does, on schedule.

“Phonics is necessary. It was never sufficient. Treat it like the whole answer and you will get a child who sounds out words and still can’t read.” The same discipline applies to AI access. The tool is necessary. It is not sufficient. Hand a child a powerful tool without training anyone to use it well, in a school without a policy governing it, and you have not modernized education. You have added a variable.

Key Takeaways:

Adoption is running far ahead of readiness: 58% of school leaders report scaling AI, but 77% of students and 53% of educators have received no formal AI training, per Microsoft’s 2026 global survey of 3,345 respondents.

The equity gap is measurable and compounding: Teachers in high-poverty schools are nearly half as likely to have AI training as teachers in low-poverty schools (39% vs 67%), according to RAND — precisely where struggling learners are concentrated.

The demand for training already exists: 66% of educators want monthly or quarterly AI guidance from their institutions; the institutions have not delivered it.

The Four Questions Worth Asking at Your School This Week

If your child’s school added AI tools this year, four questions separate a school using AI well from one using it because it felt like progress. Has the school provided structured, ongoing AI training to teachers — not a one-hour orientation, but regular formal guidance? Does the school have a written policy on how students use AI and what safeguards are in place? Are teachers using AI to sharpen their instruction, or to reduce their involvement? And is your child learning how AI produces its outputs — how to evaluate its claims, spot its errors, and push back when it gets something wrong — or only how to operate it?

The schools most likely to answer yes to all four are, per the RAND data, the ones that were already better resourced. That is a problem worth naming to your principal, your school board, and — if your child already carries a heavier learning load — your IEP or 504 team. Whether an AI tool is helping or adding noise is a legitimate agenda item in any educational planning meeting, and you do not need a credential to put it on the table.

The will, at least, is not absent. Eighty-seven percent of educators and education leaders agree that AI literacy is important for students’ futures, and 79% of students agree. What the Microsoft report documents clearly is the gap between that will and the institutional follow-through. That gap closes faster when parents ask the question out loud.

Parents built the first protective layer around their children in every prior wave of “the future is here” education reform — and that instinct was right then, too. Nobody will ever advocate for your child as hard as you will. The obstacle is not AI. The obstacle is the same one it has been in every reform cycle: a system that adopts a powerful tool faster than it trains the people holding it, then counts adoption statistics as success before checking whether the children at the bottom of the income ladder are getting the same prepared teacher as the children at the top. One question changes everything here — not “does your school use AI” but “have your child’s teachers been trained to use it well.” Ask it this week. The Brain Bloom System builds exactly the multi-system cognitive foundation — working memory, processing speed, critical evaluation — that makes a child a resilient, discerning user of any tool, AI included. Learn how it works.