92% of Students Use AI, But 77% Have Never Been Taught How
Microsoft's 2026 AI in Education Report reveals a striking paradox: 92% of students are already using AI for school, yet 77% have never received any formal AI training. Behind these numbers lies a deeper question — are we teaching kids to use AI as an 'answer machine' or a 'thinking partner'?
> Source: Microsoft AI in Education Report 2026 (Released June 24, 2026)
One set of numbers made me stop and think for a long time.
In the third edition of its AI in Education Report, released in June 2026, Microsoft surveyed over 45,000 respondents across 35 countries and reached this conclusion: **92% of students are already using AI for school-related tasks, yet 77% have never received any formal AI training.**
In other words, nearly every child is using a tool that no one has taught them how to use.
We've been here before. When Google appeared, teachers worried students would just "copy and paste" without thinking. But eventually, we did learn how to use search engines — how to filter information, evaluate sources, and treat the internet as a tool rather than an answer. AI presents the same challenge, just with higher stakes and faster pace.
## The Question Isn't "Whether to Use AI" — It's "What to Use AI For"
One line in the report stuck with me. Microsoft said their new AI learning tools are designed to make AI a "learning partner," not an "answer engine."
The "answer engine" logic goes like this: student asks, AI answers. The problem disappears, but so does the thinking. Students don't need to understand — they just need to copy. This turns AI into the smartest cheating tool ever built, rather than the most powerful learning tool.
The "learning partner" logic is completely different. When a student struggles, AI doesn't give the answer directly. Instead, it asks: "What do you think?" "Have you seen this word before?" "If that's true, what happens next?" Every question returns the responsibility of thinking to the student.
Microsoft's new Study and Learn Agent does exactly this — guiding students through interactive prompts and real-time feedback, but never doing the work for them. This design direction represents one of the world's largest tech companies formally pushing back against the narrative that "AI equals cheating."
## Teachers Are the Real Problem
Another set of numbers in the report deserves more attention: **53% of teachers have also never received formal AI training.**
A teacher who hasn't been taught how to use AI — how can they teach students to use it responsibly? How can they tell whether a student's work was written by them or generated by AI? How can they design lessons where AI enhances learning rather than replaces it?
This isn't the teachers' fault. The pace of AI tool development far outstrips any teacher training system's ability to keep up. The problem is that the entire education system is still waiting for a "definitive answer," while AI doesn't wait for anyone.
Notably, 87% of teachers and education leaders believe that "using AI effectively and responsibly" is critical to students' futures. Everyone knows it matters — but there's still a long distance between knowing and doing.
## Academic Integrity: Everyone's Worried, Nobody Knows What to Do
The report shows that 41% of students and 42% of teachers both rank "academic integrity" as their biggest concern about AI. Almost a tie.
This symmetry is telling. Students may worry about "getting caught using AI to cheat." Teachers may worry about "not being able to tell if students are cheating with AI at all." Both sides are worried about the same thing, but from completely different angles.
The solution isn't to ban AI — it's to redesign assessment. If an assignment can be entirely completed by AI, then that assignment was never measuring meaningful learning in the first place. The truly valuable learning outcomes are those that AI can't replace, only assist — critical thinking, creative application, judgment in real-world contexts.
## What I Thought About After Reading This
Microsoft concluded by saying they want AI tools to "earn teachers' trust." That line told me this company understands the current situation better than most: teachers haven't trusted AI yet — not because they're resistant to change, but because they haven't seen enough reason to believe AI genuinely helps learning rather than just making learning look easier.
The core issue here isn't technology — it's educational philosophy. Do we believe students need to think for themselves? If yes, then AI should be designed as a tool that guides thinking. If not, then AI is just a faster search engine.
92% of children are already using AI. The question is no longer "should we let them."
**The question is: are they using it as an "answer machine" or a "thinking partner"?**
That answer doesn't live in the AI itself. It lives in the choices made by the people who design AI — and the people who use it.
---
**Further Reading**
[Microsoft AI in Education Report 2026 (June 24, 2026)](https://news.microsoft.com/source/2026/06/24/microsofts-new-ai-in-education-report-highlights-widespread-adoption-and-increasing-demand-for-support/)
One set of numbers made me stop and think for a long time.
In the third edition of its AI in Education Report, released in June 2026, Microsoft surveyed over 45,000 respondents across 35 countries and reached this conclusion: **92% of students are already using AI for school-related tasks, yet 77% have never received any formal AI training.**
In other words, nearly every child is using a tool that no one has taught them how to use.
We've been here before. When Google appeared, teachers worried students would just "copy and paste" without thinking. But eventually, we did learn how to use search engines — how to filter information, evaluate sources, and treat the internet as a tool rather than an answer. AI presents the same challenge, just with higher stakes and faster pace.
## The Question Isn't "Whether to Use AI" — It's "What to Use AI For"
One line in the report stuck with me. Microsoft said their new AI learning tools are designed to make AI a "learning partner," not an "answer engine."
The "answer engine" logic goes like this: student asks, AI answers. The problem disappears, but so does the thinking. Students don't need to understand — they just need to copy. This turns AI into the smartest cheating tool ever built, rather than the most powerful learning tool.
The "learning partner" logic is completely different. When a student struggles, AI doesn't give the answer directly. Instead, it asks: "What do you think?" "Have you seen this word before?" "If that's true, what happens next?" Every question returns the responsibility of thinking to the student.
Microsoft's new Study and Learn Agent does exactly this — guiding students through interactive prompts and real-time feedback, but never doing the work for them. This design direction represents one of the world's largest tech companies formally pushing back against the narrative that "AI equals cheating."
## Teachers Are the Real Problem
Another set of numbers in the report deserves more attention: **53% of teachers have also never received formal AI training.**
A teacher who hasn't been taught how to use AI — how can they teach students to use it responsibly? How can they tell whether a student's work was written by them or generated by AI? How can they design lessons where AI enhances learning rather than replaces it?
This isn't the teachers' fault. The pace of AI tool development far outstrips any teacher training system's ability to keep up. The problem is that the entire education system is still waiting for a "definitive answer," while AI doesn't wait for anyone.
Notably, 87% of teachers and education leaders believe that "using AI effectively and responsibly" is critical to students' futures. Everyone knows it matters — but there's still a long distance between knowing and doing.
## Academic Integrity: Everyone's Worried, Nobody Knows What to Do
The report shows that 41% of students and 42% of teachers both rank "academic integrity" as their biggest concern about AI. Almost a tie.
This symmetry is telling. Students may worry about "getting caught using AI to cheat." Teachers may worry about "not being able to tell if students are cheating with AI at all." Both sides are worried about the same thing, but from completely different angles.
The solution isn't to ban AI — it's to redesign assessment. If an assignment can be entirely completed by AI, then that assignment was never measuring meaningful learning in the first place. The truly valuable learning outcomes are those that AI can't replace, only assist — critical thinking, creative application, judgment in real-world contexts.
## What I Thought About After Reading This
Microsoft concluded by saying they want AI tools to "earn teachers' trust." That line told me this company understands the current situation better than most: teachers haven't trusted AI yet — not because they're resistant to change, but because they haven't seen enough reason to believe AI genuinely helps learning rather than just making learning look easier.
The core issue here isn't technology — it's educational philosophy. Do we believe students need to think for themselves? If yes, then AI should be designed as a tool that guides thinking. If not, then AI is just a faster search engine.
92% of children are already using AI. The question is no longer "should we let them."
**The question is: are they using it as an "answer machine" or a "thinking partner"?**
That answer doesn't live in the AI itself. It lives in the choices made by the people who design AI — and the people who use it.
---
**Further Reading**
[Microsoft AI in Education Report 2026 (June 24, 2026)](https://news.microsoft.com/source/2026/06/24/microsofts-new-ai-in-education-report-highlights-widespread-adoption-and-increasing-demand-for-support/)