How can we help students use AI effectively and responsibly while developing those habits ourselves?
That’s a question a lot of teachers, schools and education systems are working through. There’s plenty of discussion about screens, technology and what students should or shouldn’t be using. But when we get down to classroom practice, we need something we can actually apply.
Something that helps us think about what AI contributes, what we bring to the process, and who takes responsibility for the result.
That last part matters. It was brushed over in quite a few of the early conversations about AI.
I recently heard a framework discussed on Guy Kawasaki’s Remarkable People podcast that gave me a useful way to think about this. I haven’t been able to track down the particular episode or speaker, so I can’t give them proper credit yet. But I wanted to share how I’d apply the idea to teaching and learning.
The framework is BILD. Think “build”, without the U:
- Brain
- – Invent
- – Leverage
- – Decide
What I like about it is that it puts our knowledge, experience and judgment at the centre of the process. And it works for teachers as well as students.
B: Brain — start with your own thinking
Before you open an AI tool, pause.
What do you already know? What are you trying to achieve? What questions do you have?
I keep coming back to this: your knowledge and experience matter when you’re working with AI. They help you give it useful direction and recognise whether what comes back is any good.
For teachers, that means starting with the learning purpose.
Before asking AI to create an activity, identify what your students need to understand and what successful learning would look like.
You might think:
My students can describe historical events, but they struggle to explain why they happened. I need an activity that helps them make those causal connections.
Now you have something useful to work with. You have a purpose to guide your request and a basis for judging the response.
For students, this might mean attempting a problem, drafting an explanation or writing their own questions before asking AI for help.
A simple classroom routine could be:
Before you ask AI, show your starting thinking.
That gives students something to compare, question and improve. It also gives us a chance to see what they understand independently.
I: Invent — explore the possibilities
This is where we generate ideas, consider alternatives and try different approaches.
You can do that on paper, through a conversation with a colleague, or using AI. You don’t have to involve AI in every step.
As an experienced teacher, you might already know how you want to introduce a topic. That’s fine. You’ve already done that thinking.
But if you want some alternatives, you could ask:
“Suggest three familiar situations in which students might need to persuade someone. Explain how each could support a short writing activity.”
Or you might describe an activity you’ve already tried, explain where students struggled, and ask for possible changes.
For students, this stage could involve developing research questions, exploring project formats or considering different explanations for a character’s actions.
We still want them contributing their own ideas. A useful question is:
What did AI suggest, and what did you add or change?
The suggestions are possibilities to investigate. They may be inaccurate, predictable or missing the context of your classroom. Teachers and students still need to bring their own knowledge and creativity to the task.
L: Leverage — use AI to move the work forward
This is where AI can do some of the heavy lifting.
You’ve thought about the purpose. You’ve considered the possibilities. Now you choose the support that will help you get the work done.
For a teacher, that might mean adapting an explanation, drafting practice questions or modifying an existing resource.
You may already know exactly how you want that resource changed. AI helps carry out those changes, saving you hours of putting it together.
I’ve used AI to create labelled, exploded-view illustrations for history lessons. I know what I want the image to explain. I supply the information, check the result and make changes where needed.
I couldn’t draw those images to the same standard myself. AI helps me produce a resource I can use to support a particular learning outcome.
For students, useful support might be a hint, a different explanation or practice questions delivered one at a time.
For example:
“Give me one hint to help me choose the next step. Let me try before you explain further.”
We need to teach students how to ask for that kind of help. We can’t assume they’ll arrive knowing how to use AI in ways that support their thinking.
The boundaries also need to be clear. If students are practising something they must demonstrate independently, AI assistance may be limited or unavailable. Explain that before they begin, alongside your school’s expectations about approved tools, privacy and acknowledgement.
D: Decide — make the judgment and take responsibility
This is the part I’d spend some time on.
AI can give you a polished, confident response that contains errors, missing context or invented references.
You still have to decide whether to use it.
Does the explanation hold up? Are the examples appropriate for your students? Does the activity actually assess what you intended?
You might accept a suggestion, change it or reject the whole thing. The final decision stays with you.
I was discussing an AI-assisted online activity with someone recently. The question came up: who made it, the teacher or the AI?
Well, the teacher came up with the idea. They set the requirements, reviewed the results and went through several rounds of changes to improve it.
There’s intellectual work in that process.
It’s a bit like saying a car designer didn’t design the car because they didn’t physically assemble it. The direction, expertise and decisions still matter.
And so does responsibility. If I put a resource in front of my students, I’m responsible for checking it.
The same principle applies to students submitting their work.
What could this look like in a history lesson?
Imagine students are explaining why a historical event happened.
- They begin with their notes and source material, then draft their own explanation. That’s Brain.
- 2. Where AI use is permitted, they ask it to suggest questions that might reveal gaps in their reasoning. They consider those questions and add their own. That’s Invent.
- 3. They choose one area to work on and ask for focused feedback. Perhaps they want to know whether they’ve explained the connection between a cause and an outcome clearly. That’s Leverage.
- 4. Then they return to their notes and sources, check the claims, decide which feedback to accept and revise their explanation. They acknowledge how they used AI. That’s Decide.
Three short reflection prompts can help make this visible:
- I used AI to…
- – I checked…
- – My final decision was…
Those statements give us something useful to discuss beyond whether a student used AI.
We can look at what they understood, what support they sought and how they judged the response.
Model the process yourself
We can show students what this looks like in our own work.
Take a question you drafted. Show them an AI suggestion you considered. Explain why you accepted it, changed it or rejected it.
Include something you rejected. That decision, and your explanation of it, might be the most valuable part of the demonstration.
You’re making your professional judgment visible. You’re showing students that a confident answer still needs thinking about.
You don’t need to redesign an entire unit to try this. Choose one task, explain the boundaries and put the four words on the board:
Brain. Invent. Leverage. Decide.
Then leave time to reflect.
Did using AI improve the teaching and learning here? What evidence can you see?
That’s the question I’d keep coming back to.
If you try this with your students, I’d be interested to hear what they found useful, what they questioned and what they decided to reject.
Look forward to your thoughts on this framework.
Kane