We Need AI That Helps Us Remain Makers

I noticed something strange the first time an AI generated in seconds what would previously have taken me hours.

The result was good. In some ways, better than what I might have made alone. It was fast, clean, impressive.

And yet I felt almost nothing.

There was no small discovery, no stubborn problem slowly becoming clear, no moment when the work began to feel like mine. I had the output, but I had missed the experience of arriving there.

This feels important.

We often talk about AI as if the central question is employment. Which jobs will disappear? Which skills will lose value? Which professions will survive?

These are real questions, but I think they are not deep enough.

The more unsettling possibility is that AI is beginning to automate not only work, but making itself.

For most of human history, making was not a specialist activity. It was woven into ordinary life. People made tools, shelters, clothing, food, songs, rituals, stories, furniture, gardens, and systems of care. Even when the object was practical, the process carried something else: knowledge, identity, memory, pride, belonging.

A person did not simply receive a finished world. They participated in shaping it.

We still see this instinct clearly in children. They stack blocks, draw animals that look nothing like animals, build homes from blankets, invent rules, make sounds, destroy what they built, and begin again. Nobody needs to explain the economic value of this activity. The act itself is enough.

Making seems to be one of the ways we learn that we can affect reality.

I move this piece and the structure changes. I cut here and the rhythm becomes sharper. I rewrite the sentence and the meaning becomes more honest. I try, fail, adjust, and slowly become capable.

This is agency in a very tangible form.

That may be why the arrival of generative AI feels so jarring, even when it is useful.

The machine does not only remove repetitive labour. It can remove the sequence of decisions through which a person comes to understand the work, and sometimes themselves.

The old relationship looked something like this:

effort, uncertainty, adjustment, result, pride.

The new one can become:

request, result.

This is obviously efficient. It is also psychologically different.

Meaningless friction and meaningful participation

The problem is not that struggle is inherently good. Much of human labour is exhausting, unfair, dangerous, or simply boring. We should use technology to remove that suffering.

But we need to distinguish between meaningless friction and meaningful participation.

Sanding the same surface for eight hours may be unnecessary. Choosing its shape, understanding its material, noticing where it resists, and deciding when it is finished may be the entire point.

The same applies to writing, filmmaking, software, music, architecture, and design.

A filmmaker does not only want a completed film. They become a filmmaker through the choices that produce it. A programmer does not only want functioning software. They build a mental model of the system while making it work. A designer does not only want a polished interface. They learn to see people, constraints, power, and unintended consequences through iteration.

Remove too much of that process and the object may survive while the person does not grow.

This is where the current language of AI feels limited.

Most AI products are presented as tools for eliminating work. Faster writing. Faster images. Faster code. Faster decisions. Faster everything.

The assumption is that less effort automatically means a better experience.

But not all effort is waste.

Some effort is where attachment forms. Some friction helps us notice. Some repetition builds judgment. Some uncertainty creates curiosity. Some difficulty turns an activity into a practice.

We know this in other parts of life.

People still cook when food can be delivered. They run when cars exist. They play instruments when every song is available instantly. They grow vegetables that would be cheaper to buy. They repair old objects that are not economically worth repairing.

They are not confused about efficiency.

They are looking for a relationship with the world that consumption alone cannot provide.

Making gives us that relationship.

It says: I was here. I paid attention. I changed something. This carries my choices.

This is also why handmade work may become more valuable in the age of generative abundance, not because human work is always technically superior, but because it contains evidence of attention.

A handmade chair carries decisions. A photograph taken by a person carries presence. A paragraph written slowly may carry hesitation, responsibility, and risk. These traces matter because another human can feel that someone was there.

AI can imitate many of those traces.

The harder question is whether it can replace the human need to leave them.

The goal should be greater human capability

I do not think the answer is to reject AI.

That would ignore how much possibility these systems create, especially for people who were previously excluded by cost, disability, education, language, geography, or technical complexity.

AI can help someone express an idea they could not previously communicate. It can help a small community build tools without waiting for institutional permission. It can translate, explain, prototype, repair, and open doors.

This matters.

But the design goal should not be total removal of the human.

It should be greater human capability.

There is a difference.

A system that replaces every decision may increase output while reducing agency. A system that helps a person see options, understand consequences, and make better choices can increase both capability and authorship.

One gives you the thing.

The other helps you become someone who can shape the thing.

This suggests a different direction for interaction design.

Instead of asking only, “How much work can we remove?”, we should also ask:

Which decisions help a person understand what they are making?

Where does judgment develop?

Which parts of the process create attachment?

When should the system suggest, and when should it remain quiet?

How can people inspect, challenge, and reshape what the AI proposes?

How can the tool preserve local knowledge, personal language, and cultural context rather than flattening everything into the most probable answer?

How can someone feel responsible for the result?

These are not decorative questions. They define whether AI becomes infrastructure for human freedom or another system that quietly reduces participation.

Perhaps the best AI tools will not behave like invisible factories.

They may feel more like workshops.

A workshop does not remove the maker. It surrounds them with tools, materials, references, and possibilities. It allows speed, but also attention. It supports experimentation without pretending that every choice is equal. It helps a beginner begin and an experienced person go further.

Most importantly, it keeps the person inside the process.

What happened during the time we saved?

This may also require us to reconsider what we measure.

A product team might celebrate that users now finish a task in ten seconds instead of ten minutes. But what happened during those missing nine minutes and fifty seconds?

Was it confusion and administrative waste?

Good. Remove it.

Or was it interpretation, learning, reflection, and choice?

Then perhaps we removed the wrong thing.

Efficiency is easy to count. Meaning is harder.

We can count outputs, generations, minutes saved, and tasks completed. It is much more difficult to measure whether a person feels capable, whether they understand what happened, whether they trust the result, or whether the experience helped them develop judgment.

Yet these may be the more important outcomes.

This is not only a question for creative professionals.

It affects education, where students can submit correct answers without building understanding.

It affects democracy, where citizens can receive summaries without learning how to examine claims.

It affects relationships, where messages can be generated without someone finding the words themselves.

It affects culture, where the visible surface of a place can be reproduced while its roots, tensions, and lived context disappear.

It affects software, where people can deploy systems they cannot explain.

In each case, AI can make participation easier.

It can also make participation thinner.

The difference depends on design.

What should technology help us become?

I keep returning to a simple question:

What should technology help us become?

For years, digital products were designed around convenience. Remove the step. Reduce the click. Predict the need. Make the interface disappear.

There was value in this. Many systems were needlessly difficult.

But a world without friction can also become a world without agency, where our needs are predicted, our choices are narrowed, our work is completed before we fully understand what we wanted, and we slowly become operators of systems built by someone else.

AI intensifies this possibility.

It can make us extraordinarily productive while leaving us strangely absent from our own work.

That is the risk I feel.

Not that machines will become creative and humans will stop being special.

It is that we may begin confusing possession of an output with the experience of making, and efficiency with freedom.

Making has never only been about producing objects.

It is one of the ways humans meet resistance, develop patience, express care, share knowledge, and discover that the world is not fixed.

It is how we become participants rather than spectators.

We should protect that.

The future should not force us to choose between rejecting AI and surrendering the entire creative process to it. We can build tools that remove pointless barriers while preserving meaningful decisions. Tools that expand what people can attempt without erasing why the attempt matters. Tools that help us move faster while still allowing us to notice where we are going.

The most human AI may not be the one that makes everything for us.

It may be the one that helps us remain makers.