What Is School Really For?

In the last chapter, I was thinking about what happens when AI makes creating something much easier.

When producing an article, an image or even a piece of software requires less effort, we naturally begin producing more.

But what happens when technology makes learning something easier?

Access to knowledge had already changed dramatically before AI arrived.

There was a time when you needed the right teacher, the right book, the right library or the right institution to understand something.

Then the internet changed that.

YouTube changed it further.

Today, a teenager can learn calculus from an MIT professor, guitar from a musician halfway across the world, or coding from someone who has never stepped into a classroom.

And now AI can do something more. It can explain, answer questions, adapt to someone's level of understanding and respond almost immediately.

That is an extraordinary development.

But it brings me back to a question that was already worth asking before generative AI arrived.

If knowledge can be accessed almost anywhere, what exactly is school for?

I want to start with something closer to my own experience.

What was I really learning at UCLA?

Consider an Executive MBA.

Many people entering these programmes are already experienced professionals.

They have managed teams, built businesses, led functions and made decisions with real consequences.

They are not entering a classroom because they have never heard of strategy or seen a balance sheet.

And much of the technical information taught in a business school can now be found elsewhere.

So what are people really paying for?

There are obvious reasons to pursue an EMBA, including career progression, qualifications, professional networks and access to opportunities. Those are legitimate reasons.

But when I think about my time at UCLA Anderson, some of the most valuable learning came from the people sitting around me.

Someone from healthcare would see a case differently from someone in technology.

Someone with a legal background would ask a question that had not occurred to someone in finance.

Someone who had worked in government might interpret a decision very differently from someone who had spent twenty years in sales.

We were often looking at the same information.

But we were not necessarily seeing the same problem.

I found that fascinating.

Even the design of the classroom contributed to it.

We were not simply sitting in rows watching a professor deliver information.

We could see one another.

The professor might introduce a case or challenge an assumption, but quite often another student would be the person who made me reconsider my position.

Sometimes I would enter a discussion reasonably confident about my conclusion.

And then someone would ask a question that made me realise I had overlooked something important.

The answer was not necessarily in the textbook.

It emerged through the discussion.

The conversation was part of the education.

Information and development

I keep coming back to a simple distinction.

Information can be delivered. Development has to happen.

You can read about leadership.

Then you can find yourself trying to lead a team where nobody agrees with you.

You can study negotiation.

Then you can try persuading someone who has absolutely no reason to accept your proposal.

You can learn the principles of communication.

Then you can stand in front of a room and discover that what sounded perfectly clear in your head does not sound clear when spoken aloud.

You can understand collaboration intellectually.

Then you can work with someone whose priorities and working style are completely different from yours.

The information matters.

But the experience develops something else.

One professor made an observation that stayed with me: getting work done through machines is often easier than getting work done through people.

I have thought about that quite a bit as AI has become more capable.

We call communication, collaboration, influence, judgment and empathy soft skills.

Yet these are often among the hardest capabilities to develop.

You do not become good at working with people simply because someone explains how people work.

You have to work with them.

Sometimes unsuccessfully.

And then try again.

That is when the distinction between information and development becomes real.

What about children?

The same question becomes even more important when I think about my daughter.

Imagine an AI tutor that knows exactly what a child understands.

It can explain a difficult concept in several ways.

It never gets tired of questions.

It adjusts the difficulty.

It provides immediate feedback.

It can help a child who is struggling without making her feel embarrassed in front of others.

And perhaps it can bring good individual support to children who would otherwise never receive it.

I find that possibility exciting.

But would it be a complete education?

A child still has to deal with other children.

She has to discover what happens when someone disagrees with her.

She has to learn how to lose.

How to work with people she did not choose.

How to speak when she is nervous.

How to respond when someone else is better at something.

How to help someone who is struggling.

How to create something that others may not like.

And sometimes, how to fail and return the next day.

When I look at education this way, sport, music, art, dance, drama, competitions and group projects begin to look less like activities around the edge of learning.

They are part of the learning.

The gym analogy

I find a gym a useful way to think about this.

A gym does not exist because information about exercise is scarce.

Anyone can find out what a push-up is.

There are thousands of videos explaining how to get stronger.

But watching those videos does not make you stronger.

You still have to do the exercise.

Education contains many equivalents of lifting the weight.

A good AI tutor could help someone learn more effectively.

But if it does the work the student needs to develop through, a better result may conceal weaker learning.

I explored that concern in Chapter 4, Output, Not Answer.

Here, the question is broader.

Schools should not simply compete with AI at delivering explanations and producing answers.

They should become better at creating the conditions in which children actually develop.

But the more I think about this, the more I realise that these experiences do not necessarily require a traditional classroom.

People develop confidence, judgment and the ability to work with others through sport, work, volunteering, family and life itself.

And simply attending school does not guarantee that a child develops these qualities.

A classroom can be passive.

A group project can be badly designed.

A school can reward grades while neglecting curiosity.

A university can charge a great deal of money without providing much that students could not have obtained elsewhere.

That makes me question something in my own argument.

If development is what matters, should we really be defending the traditional school?

Or should we be asking how schools can create more opportunities for development than they do today?

AI might actually help with that.

If technology can support individual learning, perhaps teachers can spend more time on discussion, interaction and the aspects of education that are difficult to deliver through a screen.

That is a possibility worth exploring, rather than assuming every existing classroom activity needs to be preserved.

What should change?

I would not want schools to respond by simply adding another class where students learn which AI tools to use.

The tools will change.

I would rather see schools become more deliberate about the capabilities they are trying to develop.

That could mean more discussion, not less.

More opportunities to explain and defend a position.

More work with people who have different perspectives.

More projects where the process matters as much as the final result.

More attention to how students respond when they are wrong.

And more opportunities to perform, create, collaborate and deal with situations that do not have a single correct answer.

The academic subjects remain important.

I am not suggesting that children no longer need mathematics, science, language or history because AI can explain them.

Without foundational knowledge, it becomes much harder to recognise when an explanation is incomplete or wrong.

But perhaps we should pay more attention to what students can actually do with that knowledge.

Not just:

What answer did you produce?

But:

How did you get there?

Can you explain it?

Can you defend it?

What did you learn when something went wrong?

Can you apply the idea in a situation you have not seen before?

And can you work through it with other people?

None of these are new educational ideas.

Good teachers have been doing many of them for years.

AI may simply make their importance more visible.

Where I have landed

When I think about my UCLA experience, some of the things I remember most clearly are not the concepts I learned in a lecture.

They are the conversations.

The people who challenged me.

The situations where I had to reconsider an assumption.

The times when I tried to explain my position and discovered that I had not thought it through as well as I believed.

Those experiences changed something in me.

And when I think about my daughter's education, I hope she has plenty of those experiences too.

Of course, I want her to learn mathematics, science, language, history and everything else that helps her understand the world.

But I also want her to become curious, think independently, work with people who are different from her, develop confidence and discover what she is capable of.

AI may become a remarkable resource in that journey.

And that is something to welcome.

But having access to knowledge and developing the ability to use it are two different things.

If AI makes information easier to access and explanations easier to obtain, schools should take advantage of that.

And perhaps use the opportunity to examine what they are spending their time doing.

Because what I remember most from my education was not always the information I received.

It was what happened to me through the experience.

Information can be delivered. Development has to happen.

And perhaps that is the question schools should be asking themselves as AI becomes part of education.

Not simply how to teach more efficiently, but what kind of person their students are becoming.