Lately I’ve been thinking that most things about AI boil down to two key questions: Is it a good use of the technology? And is it a good use of your brain?
On the face of it, those two questions are pretty simple yes-no answers, but in the grey areas between yes and no there’s a lot of nuance. I think we also need to contextualise them for this person, on this task, at this moment. And in that, the answers become a lot more complex.
I’ve used this framing recently in a few professional development sessions, and it’s been well received as a way to talk about student use and the use of AI by educators and professionals.
In this article, I’ll explore each question in turn, and think about ways that these two simple-seeming questions can help us to understand the increasingly complex ways people interact with AI.
Is this a good use of the technology?
To understand whether something is a “good use” of AI first requires a solid understanding of what the technology actually is. And that’s trickier than it might seem. “AI” is not one thing, and although the public eye has been drawn to ChatGPT, even so-called generative AI is much broader than just chatbots.
Earlier this year, I wrote a series of posts arguing that GenAI is more of a Swiss Army knife for computers than a text-in, text-out chatbot. The language capabilities of ChatGPT might have been what first caught the public’s attention, but the technologies themselves are far more interesting in the ways they can interact with code and software. So when we say “is this a good use of the technology,” we have to understand that the technology is not just a chatbot that can summarise, write emails, or produce essays.
In fact, the writing capabilities of large language models like ChatGPT and Claude are, I think, some of the least interesting capabilities.
By now, we’ve all encountered AI slop in the wild, where people carelessly overuse AI to produce writing. If you’re a teacher, you’re probably sick of seeing this kind of use by students, and you might immediately think that using a language model to write an essay is an appalling use of the technology. And, by and large, you’d be right.
But what if a student records chunks of their essays as voice memo, transcribes them using another kind of AI like Whisper, and then uses a large language model-based application to tidy up the writing up? Would that be an acceptable use of technology?
Or what if a student hammered out an essay, furiously banging on the keys like the typist in Leroy Anderson’s ‘The Typewriter Song’, and then used an AI like Copilot in Microsoft Word to tidy up the formatting; all those fiddly subheading styles that make navigation much easier?
If a student used a technology like the free version of ChatGPT to tackle a complex math question, they might be throwing a predictive text model at a problem out of its reach: ChatGPT might “guess” the answer and prove hopeless. But if a student used a more capable AI model that has the capacity to write and execute code, or perhaps an AI as an extension in a piece of spreadsheet software, then it could absolutely complete complex computation. A sufficiently capable AI could in essence build itself a calculator to do the job.
To answer the question, “is this a good use of technology”, you need to be aware of how far the technology has come in the last few years. And as I wrote about in that IYKYK series, it’s really hard, because the developers sort of want you to treat it like a fancy chatbot.
To complicate things further, the question changes across all of the contexts that I mentioned earlier: for this person, on this task, at this moment.
Let’s come back to students writing essays. One student might benefit enormously from having access to an AI model that can transcribe their handwritten notes and help them to organise their thoughts. Another student might benefit from a similar platform that takes voice memos as the main input. A third student might benefit more from just putting pen to paper. And a fourth in the same class might throw the essay topic into ChatGPT and try to generate a response wholesale.
Three of the students are using the technology, one is deliberately choosing not to, but in all four cases, whether the use of the technology is “good” varies greatly.
Is it a good use of your brain?
Again, this is a much more complex question than it seems, and even more personal than the previous question. The previous question required an understanding of the capabilities of the technology, its strengths and limitations, and what it’s suited for, with a clear-eyed understanding that AIs are increasingly more capable across a range of tasks than many people think, including mathematics, computation and communication. This question, however, needs us to look inwards at our own strengths and limitations.
“Is it a good use of your brain” needs you to ask things like: is this a skill that I need to develop for its own sake, or can I afford to outsource it? Do I have enough existing knowledge and expertise to use the technology well in this task? Am I outsourcing or just offloading, getting the AI to do work that I should really be doing myself?
These questions are reflective and metacognitive, and being reflective and “thinking about thinking” is hard work. So, there are questions of effort and capacity involved here too. We all have fluctuating capacity. Perhaps one day a certain use of the technology might be completely acceptable, and on another day when we’re tired or anxious or feeling impatient, the exact same use of the technology might be a terrible idea.
“Is this a good use of your brain” becomes an infinitely more complex question when you accept that your brain is not a singular, static thing. Your brain on a Tuesday is an entirely different animal to your brain on a Friday. Your brain at 4pm might be totally different to your brain at 7am. Like “is this a good use of the technology,” the brain question is highly contextual from person to person and moment to moment. And who can say on behalf of someone else, this is good for you and this is not?
There will be some times when a teacher could stress to a student, you really shouldn’t use AI for this task, because if you do, you’ll offload too much of the work and you won’t learn it for yourself. But can we guarantee that that’s true? Do we know how the student plans to use the technology? Have we successfully answered the first question?
If I were back in the English classroom, I might confidently say, “I’d like you to write the first draft for yourself because I want to see your ideas on the page.” But do I know that all of the students in front of me have the capacity to go directly from brain to page? Is there some potential there that the technology could intercede in a helpful way? And if there is, do I make that available to all of the students, or go case by case, brain by brain?
If I reflect on my own use of the technology, and if I’m honest with myself, it’s incredibly tempting to use AI in ways which are not good for my brain. My brain is something I’ve written about elsewhere on the blog, including why the constant, unwavering stream of AI attention might be particularly attractive and particularly dangerous, running the risk of addictive behaviours and overuse. But of course that’s my brain, and it might not be yours.
Simple questions for a world of complexity.
For a long while now, I’ve been grappling with this tension between trying to simplify AI and make it easier to use and to understand for both educators and students, and acknowledging its complexity. We tried something like that with the various versions of the AI Assessment Scale. And in a pair of articles earlier this year, Beyond Scales – In Theory and In Practice, I wrote about where I thought we’d missed the mark and where others had started to bridge the gap. For example, one of the problems with the AIAS has always been delineating the types of AI use: planning, drafting, collaborating, and so on.
In one sense, I think it’s important that we put deliberate boxes around types of AI use, because it helps students to understand what “good” looks like and it helps educators with a frame of reference to the kind of work that they’re familiar with. But if we view the AI Assessment Scale (or any other framework) through the lens of these two questions, it starts to fray at the edges.
Because if we set a task at Level 2, AI for planning, and we say “is this a good use of the technology”, then the first response might rightly be, which technology? Is the AI being used in the planning process something like Microsoft Copilot provided at the most basic level of enterprise licensing? Or is it ChatGPT 5.6 running Deep Research? Is it an AI-powered brainstorming tool built into an edtech platform like Magic School or Google’s Notebook LM? Or is it a student using Claude Code to turn a photograph of their own brainstorm in a notebook into a digital mind map using mermaid.js?
Planning, brainstorming, and taking notes might be an appropriate place for some students might be a good use of their brain, like my own experience of heavily relying on voice memos and transcription to think my way through articles like this one. Or, it might be a really bad idea, leading to students following an AI down infinite rabbit holes, or getting stuck on ideas that are tangential to the task at hand. One student’s brain might be wired in such a way that a particular use of AI is helpful at “Level 2”, and another student is hopeless.
I think frameworks are useful and necessary for now, while we all come to terms with the implications of these technologies. But like Phillip Dawson said recently in an RMIT podcast, I agree they have a shelf life of single-digit years. Not just because the technology is becoming more sophisticated and more increasingly woven into wearables and other systemic layers of our digital infrastructure, but because they cannot fully account for the individuality and nuance of these technologies and the people that use them.
And realistically, I don’t think there’s anything that can.
So, is this another framework?
When I was first kicking around these two questions, I wondered if perhaps they would make a suitable framework; maybe a decision tree or a flowchart. Is it a good use of the technology? Yes, no, maybe, when, why, why not? But I quickly ended up at the point where the branching decisions became so numerous and complex that they were ridiculous. In fact, here’s an attempt generated by Claude Code again using the mermaid.js mindmap approach shown earlier:
We’re long past the point where simple frameworks and flowcharts can hope to capture the complexities of AI.
So, if these two questions aren’t a framework, what are they? A discussion starter? A challenge, maybe? At the very least, I’d encourage you to ask yourself these two questions and see if you can answer them honestly for your own use. The next time you’re tempted to crack the lid on ChatGPT or your AI platform of choice, ask yourself: do I really understand what I’m doing here with this technology? Do I know this technology well enough to know that this is the best way to use it in this situation? Is there something that this technology can do that I’m missing? Is there another technology that would do a better job?
And then, am I thinking this through carefully enough? Is this a good use of my knowledge and expertise, or do I have gaps that I should be filling before I turn to AI? Am I using AI because I’m tired or bored or under pressure to complete a job to a deadline? Is this use of AI encouraging any kind of growth or learning, or am I just going through the motions here? Does it matter if I am going through the motions?
Once you’ve asked yourselves those questions, maybe some of that will carry forward into your assessment or curriculum design, or whatever your day job is. For me, these questions have provided a helpful check-in, a way to perhaps stay on top of some of my own brain’s less helpful tendencies, and a way to keep pushing myself further in my understanding of the technology. I hope you find them helpful, even if they are hard work.
Want to learn more about GenAI professional development and advisory services, or just have questions or comments? Get in touch:
Leave a Reply