Written by Ts Dr Suhailah Mohamed Noor | Founder, Prompt Academy
Date published: 26 Jun 2025
A Real Conversation That Hit Home
I recently had a conversation with a fellow lecturer from the Department of Mathematical and Computer Sciences. He teaches C++ to engineering diploma students and has started noticing a worrying trend: despite clear instructions not to use AI, students are still submitting AI-generated code — and pretending it’s their own.
But here’s the thing: he knows.
Not because he’s an AI expert, but because the code is not on their level.
Syntax too advanced. Logic too structured. And when asked to explain it?
Silence. Blank stares.
✅ AI-Aware, Not AI-Dependent
This got me reflecting deeply. I’m a strong advocate of AI-aware learning — I believe AI can and should be part of our classrooms. But there’s a difference between using AI to support thinking, and using AI to avoid thinking.
This is what I call responsible augmentation. It’s the heart of AI-aware education.
I’m not anti-AI. I’m just anti-empty learning.
🧠 Fundamental Knowledge Is Non-Negotiable
My colleague is absolutely right — just because a piece of code runs doesn’t mean the student understands it. AI can generate solutions that look brilliant. But if a diploma-level student can’t even explain a basic for loop, that’s not learning. That’s outsourcing your brain.
This isn’t about “catching cheaters.” It’s about recognising a cognitive mismatch between what’s being submitted and what’s truly understood.
🚨 The Real Issue: Misaligned Assessment in an AI World
Let’s be honest — students use AI because:
- Assignments often reward results, not understanding.
- They believe AI is a shortcut, not a tool.
- There’s no accountability loop post-submission.
But that changes fast when a lecturer asks:
“Boleh explain line ni kenapa
int* ptr = &x;?”
And the student suddenly realises:
You can’t fake understanding.
🛠 A Better Way: Layered AI-Aware Assessment Design
We don’t have to ban AI.
But we do need to redesign how we assess learning.
Here’s a model I strongly recommend — especially for coding, engineering, or any applied subject:
✍️ Part A: Manual Fundamentals
Students write and explain 2–3 basic problems without AI. This builds core understanding.
🤖 Part B: Supported AI Task
Students use AI to solve a more complex variation — but they must:
- Annotate or comment on the AI output
- Reflect on what the AI did, and what they learned
🗣 Part C: Spot Check
Conduct random 2-minute oral Q&A during class:
“Can you explain what this line means?”
“What would happen if I changed this value?”
If the student can’t answer → mark reduction.
No punishment. Just reinforcement of one thing: learning matters.
This structure trains:
- Integrity
- Understanding
- Accountability
- Reflective practice — the core of AI-aware education
💬 Final Thoughts: Thinking Comes First
This whole experience reminded me why AI can’t and shouldn’t replace the thinking process.
I’m not here to vilify students. They’re trying to cope — and AI feels like a lifeline. But unless we, as educators, help them build the mindset and framework to use it well, they’ll end up copying answers they don’t understand — and missing the entire point of education.
Let’s be honest:
- AI is not the problem.
- Skipping the thinking process is.
So to all educators wondering where to draw the line:
Let’s not block AI. Let’s build learners who can think, then prompt.
AI can’t replace thinking. But it can empower those who are willing to think.