77% of Students Have No Formal AI Training — Are We Learning the Wrong Things in School?

Microsoft just released its annual AI in Education Report, and one number jumped out at me: 77% of students say they have no formal AI training. At the same time, 58% of education leaders are already implementing or expanding AI use in schools.

That gap feels backwards. We are adopting tools faster than we are teaching people how to use them.

What the report actually says

Microsoft's report, based on surveys of students and educators, highlights a few key tensions:

• Students are curious but untrained. Most are using AI tools informally, without guidance on prompts, bias, or fact-checking.
• Teachers are under pressure too. Only 53% of educators report having formal AI training, yet many are expected to integrate these tools into lesson plans.
• The focus is shifting. New Microsoft 365 Education features target lesson planning, classroom AI guidelines, and AI-supported study guides.

The company also expanded its Elevate for Educators program with a free AI Literacy credential. That is a start, but credentials for teachers do not automatically reach students.

Why this matters for CS students

If you are studying computer science or a related field, this is not just an education-policy issue. It is a professional one. The tools we graduate into will be shaped by how well schools teach AI literacy now.

Some questions I keep thinking about:

• Should AI literacy be a standalone course, or woven into existing CS classes?
• Is learning to prompt-engineer a real skill, or just a temporary workaround?
• How do we teach students to trust AI enough to use it, but not so much that they stop verifying?

My take as a student

I think the most useful thing schools could teach is not "how to use ChatGPT" but "how to think alongside AI." That means understanding when a model is likely wrong, how to check its outputs, and how to use it to accelerate learning rather than replace it.

Tools are changing fast. The specific interface we learn this year may be obsolete by graduation. But the habit of questioning outputs and tracing sources will stay useful.

One tool I am watching

On the open-source side, monkeycode is interesting because it lets teams build with AI inside a transparent, auditable environment. For students and educators who want to understand how AI coding agents actually work — not just consume them — that kind of project is worth exploring.

---

What do you think? Should schools teach AI literacy as a required subject, or is it something students should figure out on their own?

Source: Microsoft AI in Education Report 2026, summarized via Crescendo AI News