There is a story about artificial intelligence that we keep hearing because it is comforting.

AI will make us more productive.

It will eliminate boring work.

It will create new industries.

It will give ordinary people access to extraordinary capabilities.

And, eventually, everyone will benefit.

I hope that story turns out to be true.

But I don’t think we should assume it will be.

Because technological revolutions have never distributed their benefits equally.

And artificial intelligence may be about to make that problem considerably worse.

The question isn’t whether AI will create wealth

It almost certainly will.

The real question is:

Who will own that wealth?

That is a much more uncomfortable question.

Imagine a company that once needed 1,000 employees to perform a particular set of tasks.

Now imagine that increasingly capable AI systems allow the same company to accomplish much of that work with 200 people.

The company becomes dramatically more productive.

Its costs fall.

Its profits rise.

Its investors celebrate.

But what happens to the 800 people who are no longer needed?

We are often told that they will simply move into new jobs.

Perhaps some will.

But that answer hides an enormous assumption: that the new jobs will appear quickly enough, that displaced workers will have the skills to perform them, and that the economic system will distribute the gains fairly.

History gives us no guarantee of any of those things.

AI doesn’t need to replace everyone

This is where I think the public conversation gets AI wrong.

People imagine a future where robots and machines replace every human worker.

That may happen eventually.

But it doesn’t have to happen for AI to transform the economy.

Suppose AI makes one highly skilled employee twice as productive.

A company might not hire two additional employees.

It might simply keep the one employee.

That means productivity increases without employment increasing at the same rate.

Now multiply that across millions of workers.

You don’t necessarily get mass unemployment overnight.

You get something subtler.

Fewer people are needed to produce more economic output.

And that changes the bargaining power of workers.

The most valuable employee may become the person who knows how to use AI

There is another possibility that deserves more attention.

AI may not eliminate entire professions.

It may eliminate the need for as many people within those professions.

A company that once employed twenty junior designers might eventually need five experienced designers using sophisticated AI tools.

A newsroom that once required dozens of people might operate with a much smaller staff.

A software company might produce more software with fewer programmers.

A small business owner might use AI to perform work that previously required accountants, marketers, designers, researchers and administrative staff.

This creates an interesting paradox.

AI can democratize access to capabilities while simultaneously concentrating economic power.

A single individual can suddenly do the work of an entire small team.

That sounds wonderful.

And it is.

But imagine what happens when the person with the AI isn’t an ordinary individual.

Imagine it is a corporation worth hundreds of billions of dollars.

It has the data.

It has the computing infrastructure.

It has the engineers.

It has the distribution network.

It has the capital.

AI becomes not merely a tool for productivity.

It becomes a force multiplier for whoever already has resources.

This is where inequality becomes dangerous

We tend to think of inequality as simply a difference in income.

But technological inequality can be much deeper.

Imagine two people.

Person A has access to the best AI systems, expensive education, high-speed internet, capital, professional networks and a company willing to invest in experimentation.

Person B has an outdated education, limited access to technology, little savings and a job that can be partially automated.

Both technically have access to “AI.”

But they are not participating in the same technological revolution.

One person uses AI to create companies.

The other watches AI transform the job market around them.

That difference matters.

Because technological progress does not automatically create social progress.

The uncomfortable possibility

There is a future in which AI makes society enormously wealthier while millions of people feel poorer.

That sounds contradictory.

It isn’t.

A country can produce more wealth while ordinary people’s share of that wealth stagnates.

A company can become more profitable while reducing its workforce.

An individual can become extraordinarily productive while another person’s skills become less valuable.

The economy can grow.

The stock market can rise.

GDP can increase.

And people can still feel that something has gone terribly wrong.

That is why the AI debate cannot be reduced to a question of whether artificial intelligence is “good” or “bad.”

The technology isn’t the real issue.

The distribution of its benefits is.

We are entering a period where intelligence itself is becoming cheaper

For most of human history, intelligence was expensive.

If you wanted someone to write something, you needed a writer.

If you wanted legal research, you needed lawyers or researchers.

If you wanted software, you needed programmers.

If you wanted a marketing campaign, you needed marketers.

If you wanted tutoring, you needed teachers.

AI is changing the economics of many of these activities.

It is making certain forms of cognitive work dramatically cheaper.

That is an extraordinary development.

But when something becomes cheaper, the people whose livelihoods depended on its scarcity face a difficult question:

What happens when the thing you were paid for is no longer scarce?

This is not a theoretical question.

It is already beginning to appear across industries.

And the answer cannot simply be “learn to code.”

Not everyone can become an AI engineer.

Not everyone wants to.

And eventually, even some of the work performed by programmers will itself be assisted—or replaced—by increasingly capable AI.

Telling everyone to acquire another technical skill is not an economic strategy.

It is postponing the question.

So what should we actually do?

First, we need to stop pretending that technological progress automatically produces broadly shared prosperity.

It doesn’t.

Prosperity has to be designed, negotiated and distributed.

That means governments will eventually have to rethink education, taxation, labor policy and social protection.

Businesses will have to decide whether AI is merely a mechanism for reducing headcount or a tool for making their existing employees dramatically more capable.

And individuals have to recognize that the world of work is changing whether they are ready or not.

But there is something even more important.

We need to redefine what we consider valuable.

If machines become extraordinarily good at producing information, perhaps human value will increasingly come from judgment.

Trust.

Leadership.

Creativity.

Relationships.

Responsibility.

The ability to understand other human beings.

The ability to decide what should be done—not merely how to do it.

Those qualities may become more valuable precisely because machines are becoming so good at everything else.

The greatest mistake would be to panic

I don’t believe the answer is to fear AI.

And I don’t believe the answer is to worship it either.

Both reactions are intellectually lazy.

We should neither assume that AI will destroy humanity nor blindly assume that it will create paradise.

We should ask better questions.

Who owns the machines?

Who controls the infrastructure?

Who receives the productivity gains?

Who bears the cost of displacement?

Who gets educated?

Who gets left behind?

And perhaps most importantly:

What kind of society do we want to build with this technology?

Because AI will not answer that question for us.

Humans will.

And if we leave that decision entirely to corporations, markets and whoever happens to control the most powerful systems, we shouldn’t be surprised if the future primarily serves their interests.

The AI revolution is coming.

The question isn’t whether it will make us richer.

It probably will.

The question is whether it will make most of us richer.

Those are two very different things.

And we should start talking about the difference now, before the answer is decided for us.