There are few conversations in education at the moment that seem to move as quickly as the one about artificial intelligence.
In a relatively short period of time, we have moved from questions about whether students should use AI at all, to increasingly complex conversations about how, when and why they might use it. Along the way, concerns about cheating and academic integrity have been joined by questions about creativity, critical thinking, metacognition and whether, by making some things easier, we might inadvertently make learning poorer.
These are reasonable concerns. There are things about learning that are worth protecting.
The struggle to understand something. The process of forming an idea. Learning how to question, create, critique, revise and persist. Developing the ability to recognise when we understand something and, perhaps more importantly, when we don't. If AI removes too much of that process, there is a risk that students arrive at an answer without experiencing the learning required to get there.
There is another side to this. Schools have always had a responsibility not only to educate students for the present, but to prepare them for what comes next. And while we cannot know precisely what the future of work will look like, it seems increasingly difficult to imagine one in which AI does not play some part.
This creates an uncomfortable tension. In trying to preserve an authentic learning experience today, could we inadvertently deprive students of authentic opportunities for their future?
Protection and preparation
The temptation is to reduce this tension to a question of how much AI students should be allowed to use. Perhaps that is the wrong question. A more useful one might be:
“When does AI strengthen the learning we value,
and when does it substitute for it?”
The distinction matters. Using AI to avoid thinking is quite different from using it to extend thinking. Generating an answer is different from interrogating one. Asking a tool to do the creative work is different from using it to challenge an idea, explore alternatives or provide another perspective.
The educational opportunity may therefore be less about teaching students how to use particular AI tools and more about developing the judgement to decide when they should use them at all.
In my time as an ICT Integrator, I’ve always advocated for “The right tool for the right job”. That judgement will be increasingly important because the tools themselves will continue to change.
For those responsible for strategy and leadership in schools, this presents a particular challenge. Over the past three years, it has sometimes felt as though the moment an AI policy or guideline is established, the technology moves again. New capabilities appear. Platforms incorporate features that previously existed only in specialist tools. What seemed appropriate six months earlier suddenly requires reconsideration.
It is difficult to create certainty around something that is inherently uncertain. Perhaps, then, our policies need to be anchored less to individual technologies and more to the principles we are trying to uphold.
What do we believe good learning looks like? What capabilities do we want students to develop? What should remain distinctly human? Where might AI genuinely enhance learning? And where might its use diminish the very thing we are trying to achieve?
These questions are harder to turn into a neat policy document. But they may also endure longer than rules written around today's technology.
Guardrails are not the destination
None of this suggests that schools should simply open the gates. Guardrails matter. Age, privacy, safety, academic integrity and developmental readiness all require careful consideration. There will be contexts where limiting or removing access to AI is entirely appropriate.
But guardrails mostly tell us where the boundaries are. They do not necessarily show us what good practice looks like. That may be the next important step.
Alongside policies and guidelines, we need examples of meaningful practice: case studies that demonstrate where AI has genuinely improved learning, where it has helped a student think differently, or where choosing not to use AI has been equally important. Students need these examples. So do teachers. And increasingly, so do parents.
There is an understandable expectation that schools will prepare young people for a future shaped by AI. Equally understandable is the concern that introducing it too readily could weaken some of the capabilities young people will need most.
Both can be true. Perhaps the response is neither enthusiastic adoption nor cautious avoidance, but considered use.
We do not know exactly which jobs today's students will eventually enter. Some existing roles will change. Some may disappear. Others, including ones we cannot yet imagine, will emerge. That uncertainty need not be approached only with fear.
If anything, it reinforces the importance of developing young people who can think critically, create, adapt, question, learn and exercise judgement.
AI should not give us a reason to value those capabilities less. It should make us value them more.
So perhaps the challenge for education is not to choose between protecting students from AI and preparing them for it.
It is to protect the experiences through which meaningful learning occurs, while giving young people enough opportunity to develop thoughtful, critical and purposeful ways of working with the technologies that will increasingly surround them.
The question is not simply how much AI we allow into education. It is what we want to preserve, what we are prepared to change, and why.