Issue 139

The two most common arguments I hear about integrating AI into the classroom go something like this:

1. Not teaching students how to use AI may leave them at a disadvantage in the future job market. Companies are expecting their entry-level employees to know how to leverage it. As a result, we have a responsibility to teach it.
2. AI is not going anywhere. Students are already using it. We can’t teach as though they aren’t.

It might be the case that AI proficiency is needed in the future of work. It might also be the case that our students will use both agentic and generative AI daily. In this essay I’ll grant both of these points. Even so, I will argue that some classrooms should remain free of AI integration.

In fact, the two arguments above make it all the more important to preserve some AI-free spaces. Schools should intentionally protect places where students can practise organic thought – free from the constant influence of the technologies that increasingly mediate the rest of their lives. In what follows I’ll make a case for why – a case that has absolutely nothing to do with cheating.

1. We should mitigate AI co-dependency

Knowing how to use a tool is different from using it for everything. It can be the case that a student needs to learn proficiency with a tool to perform certain tasks. This does not mean the tool must be integrated into all, most, or even a majority of their academic work.

Advocates who push for that are not pushing for deeper learning. They are pushing – often unintentionally – for systematic AI co-dependency.

This entails fostering an environment where students become so wedded to AI tools (either through intentional integration or simple negligence) that they immediately reach for their phones to think through something with Claude or ChatGPT.

Sadly, this is already the case for many of our students.

• Discussion questions in small groups? Students are feeding it into their favourite LLM before even attempting to read the questions themselves.
• Problem with a friend? Run it by AI.
• Difficult email? Run it by AI.
• Social media caption? Run it by AI.

These are the little tasks that make up our lives. If we give all of this away, we slowly give away our capacity for judgment because we are no longer capable of sitting with the discomfort that comes with a small roadblock or a moment of unclarity.

The wisdom needed to make good decisions requires practice: reasoning through situations, exercising moral imagination, and discerning how principles apply in concrete circumstances. If students outsource the early stages of reasoning to AI, they never build the mental muscles required for judgment.

The classroom should be the place where questions come alive because students have the space to pursue them.

None of these capacities can develop if we continually outsource opportunities to brainstorm, to wrestle with the unique context a situation calls for, or to discern how a principle ought to be applied. AI will be called upon for all of it if we aren’t careful about enforcing some limits.

We owe it to our students to let them know:

• You can think through hard things.
• You can form your own thoughts.
• You don’t need an immediate answer.
• You don’t need a polished answer. The only thing you need is your answer.

How do we give that message weight? How do we back it up? We give them the space to do it. We don’t simply tell students to think for themselves and then send them home where their agency is effectively thwarted – to an enviroment where everything nudges them to outsource that thinking.

In other words, our classrooms should be spaces that help students resist the co-dependency that big tech desperately hopes they develop.

2. Classrooms should be retreats for slowness

Think about the conditions that allowed for your deepest insights.

Did they include a TV blasting? A frantic dash before some deadline? The latest tweet from an influencer? The blatant sycophancy of your favourite chatbot?

No? I’m shocked.

When we have the time and the space to question, reflect and stop – just to stop and listen – these are the conditions under which our souls come alive with curiosity and intrigue.

Learning has an ecology. Certain environments nurture attention, curiosity and intellectual courage. Others erode them. If students should be afforded one place where they can stop, question and reflect, it should be the classroom. We should give them intentionally crafted time for stopping and thinking.

They [AI tutors] offer the allure of intellectual progress without the discomfort that genuine inquiry often requires.

The classroom should be the place where questions come alive because students have the space to pursue them. Where they spend afternoons in conversation teasing out ideas. Where they stop the clicking and scrolling and beeping, listen to words spoken by other humans, and respond with words of their own.

Why does this require resisting the integration of AI into every learning activity? Because AI encourages the exact opposite moral tempo. It is the natural extension of our convenience culture: learning is not valued as a search for truth but as a process judged by outputs. It must be measured. It must be quantified. It must produce – preferably immediately. LLMs, for example, promise faster production. AI tutors promise frictionless learning and efficient measurement of each learner’s precise needs so their learning gaps can be closed ever more quickly. They offer the allure of intellectual progress without the discomfort that genuine inquiry often requires.

And the lack of friction is precisely what modern people have come to equate with the good life.

It is not. And why are we even rushing?

3. We should teach and model critical engagement with AI

Critical thinking means asking good questions. Demanding compelling reasons. Challenging weak evidence. Exercising sound judgment.

Teachers modelling critical engagement with AI are asking:

• Does AI support actually deepen learning?
• Are students strong enough to resist outsourcing their thinking outside the classroom?
• Is it wise to integrate AI into every cognitive task when the creators of these tools have powerful incentives to maximise dependency?

Like all questions worth asking these deserve rigorous debate. But debate requires spaces where the technology itself is not already taken for granted – spaces where it is not merely accepted as normative.

AI-free classrooms keep these questions alive in real, tangible ways. They model something increasingly rare in a technological culture: the willingness to pause, to question and to critically examine the tools that promise to solve our problems.

Case closed

An AI-free classroom is not a rejection of technology. Though AI may shape the future of work, it does not merit our total surrender. We should remain committed to forming human beings who are capable of thinking without it. Moreover, we want to form people who can question the designs that are sold as necessary, the motives that are sold as genuine, and the structures that are sold as inevitable and beyond our control. They are not always necessary, not always genuine, and most certainly not inevitable.

AI-free classrooms help our students step back and have the courage to do something about their answers.

 

Read Lily's full Substack article here.