I was helping my niece with her geography homework last weekend. She was researching coastal erosion, and within seconds, she had a perfectly structured, five-paragraph essay summarising the topic. It was fluent, factually correct, and utterly devoid of any spark of her own curiosity. She hadn't written a word of it. An AI did.
The Tool vs. The Crutch
It got me thinking. That AI is an incredible tool. A student could use it to generate a basic structure, then dive into each point, challenge the assumptions, and find their own examples. They could ask it to play devil's advocate. But that's not what happened. It was a copy-paste job. The thinking stopped the moment she got a 'good enough' answer.
And honestly? I get it. When you're faced with a blank page and a deadline, the path of least resistance is terribly appealing. I'm sceptical of any claim that tech will magically fix engagement. It's all in how you use it. A power drill in my hands might hang a picture. In a carpenter's, it builds furniture. Same tool, wildly different outcomes.
The line is so thin.
Passive Consumption vs. Active Dialogue
So what's the difference? It boils down to a simple question: is the student having a dialogue, or are they just collecting a monologue?
If you treat AI like a magic oracle that spits out finished work, you learn nothing. You remain passive. But if you treat it as a sparring partner - something to ask "why?" and "how?" and "could this be wrong?" - then you're actively thinking. You're challenged. The tool stimulates you by forcing you to refine your questions, not just accept its answers.
My mate who teaches design told me he now sets briefs where students must use an AI image generator, but then have to critically analyse and manually sketch three improvements to the output. Suddenly, they're not just consumers; they're editors, critics, and creators. That's the shift.
Building Tools for Thinking, Not Just Doing
This mindset is something I try to bake into the tools I build. Take Timestamp Bookmarks for YouTube. On the surface, it's just for saving points in a video. But the goal isn't to passively watch. It's to actively engage.
You're watching a long programming tutorial. Instead of zoning out, you hit Alt+S when the lecturer introduces a tricky concept. You label it "Recursion example - need to practise." Later, you jump back, not to rewatch mindlessly, but to actively tackle that specific problem. The tool doesn't learn for you. It structures the material so you can learn better. It turns a passive lecture into an interactive resource.
That's the kind of assistive tech that matters.
Asking Better Questions
Maybe the real skill we need to teach now isn't how to find answers. Anyone can do that. It's how to ask better questions. How to probe, to doubt, to synthesise. To use the AI's output as a starting point for your own ideas, not the final destination.
Otherwise, we're just training a generation to paint by numbers. And the world needs more than that. It needs people who can look at the blank canvas and see something new.
So, is AI good or bad for students? Honestly, I don't think that's the right question anymore. It's here. It's a fact. The question is what we do with it. Do we let it do the thinking, or do we use it to think even harder?
I know which one I'm hoping for.