The Jobs Nobody Can Describe Yet

Everybody seems to know which jobs artificial intelligence is going to eliminate.

Accountants, programmers, customer service representatives, administrative assistants, and quite a few others are apparently on the endangered species list. Add robots to the equation, and we can start worrying about warehouse workers, delivery drivers, factory employees, and perhaps the fellow who mows my lawn.

The predictions are everywhere, generally accompanied by rather alarming numbers.

But I have a question.

Where are all the new jobs going to come from?

Not just the jobs required to build data centers, manufacture robots, or maintain AI systems. We already know something about those.

I’m interested in the businesses, industries, products, and services that don’t exist yet but may become possible because of these technologies.

And I suspect we’re overlooking something rather important.

Nobody Predicted the Influencer

Consider one of the more remarkable business developments of the internet era: influencers.

Not long ago, explaining that you intended to make a living by posting videos of yourself on the internet might have raised serious questions about your employment prospects.

Today, influencers can earn substantial incomes by attracting audiences and persuading people to buy products, visit destinations, or adopt particular lifestyles.

Goldman Sachs estimated that the creator economy could approach $480 billion by 2027.

Obviously, not everyone posting videos of their cat is getting rich. Most aren’t. But an entirely new commercial ecosystem emerged around something that once looked more like a hobby than a business.

And it didn’t stop with individuals.

Successful creators developed businesses that employ production crews, editors, marketing specialists, financial managers, and others. Advertising agencies and technology platforms developed services specifically for this market.

An individual activity evolved into something approaching an industry.

How many other business models are waiting to be discovered?

What the Internet Really Changed

When we discuss technological revolutions, we frequently point to the automobile replacing the horse and creating entire industries in the process.

That’s a useful historical example, but I think the internet provides a more interesting comparison.

The internet didn’t fundamentally change how most physical products were manufactured. We still needed factories, equipment, materials, and workers.

What it revolutionized was how products were marketed, sold, and distributed.

Before the internet, a small retailer’s potential customers were largely determined by geography. A specialty store in Reno might attract customers throughout the region, but reaching somebody in London or Tokyo was a considerable undertaking.

The internet changed that.

Suddenly, a small business could reach customers virtually anywhere. Market size was no longer determined primarily by how far somebody was willing to drive.

Of course, selling internationally still involved shipping, payments, regulations, and other complications. But an enormous geographical barrier had been reduced.

That expansion created businesses and jobs nobody could have accurately predicted when the internet first appeared.

What if AI does something similar, but on a much broader scale?

What Happens to Manufacturing?

This is where things become particularly interesting.

Much of the innovation in manufacturing over recent decades has focused on productivity, automation, quality, and cost reduction.

Those improvements generally make existing production processes more efficient.

But what if AI changes what can economically be produced in the first place?

Imagine an individual developing a specialized product for a relatively small market.

Historically, turning that idea into something customers could purchase might require engineers, designers, prototypes, manufacturing facilities, substantial financing, and distribution arrangements.

Many worthwhile ideas never became businesses because the cost of getting started was simply too high.

AI-assisted design, simulation, robotic manufacturing, and additive manufacturing could change that equation.

Products that once required large production runs might become economical in smaller quantities. Specialized products might be designed for individual customers. Manufacturing might move closer to the markets being served.

Some of this is already happening.

The larger possibility is that manufacturing could become accessible to entrepreneurs who previously couldn’t afford to enter the market.

Of course, producing more goods with fewer workers doesn’t automatically create employment.

But creating thousands of businesses that otherwise wouldn’t exist is a different economic proposition.

And some of those businesses may eventually become substantial enterprises.

Beyond Geography

There’s another possibility that I find even more intriguing.

The internet reduced geographical barriers to commerce.

Could AI reduce language, cultural, social, and even some political and economic barriers?

Consider a small American company attempting to sell its products in Japan.

Language is one obstacle. Understanding local purchasing preferences is another. Regulatory requirements, technical documentation, customer support, and business practices introduce additional complications.

Large corporations can afford specialized departments to handle these problems.

Small businesses usually cannot.

But AI could make many of those capabilities accessible at relatively modest cost.

A small company might communicate with customers in multiple languages, adapt its products to local markets, understand foreign regulations, and provide international customer support without establishing offices around the world.

The opportunity isn’t simply to reduce costs.

It’s to make markets accessible that previously weren’t economically practical to pursue.

And that works in both directions. Entrepreneurs in smaller or less developed economies could gain access to capabilities and customers that were previously beyond their reach.

Political and social barriers are another matter. Technology cannot simply erase national interests, trade restrictions, distrust, or cultural differences.

But it might reduce misunderstandings and make cooperation easier.

Which raises an interesting question about AI safety.

Much of the public discussion focuses on the dangers of powerful artificial intelligence. Those concerns deserve attention.

Yet the standards, safeguards, accountability, and international agreements developed to manage those dangers could also help establish the trust necessary for AI-enabled commerce.

Perhaps the same frameworks designed to prevent irresponsible uses of AI could help make responsible international business easier.

That wouldn’t eliminate political disagreements. But it might provide a more dependable foundation for economic cooperation.

From Individuals to Industries

I suspect many of the new businesses will begin with individuals or very small organizations.

Someone discovers a way to provide a service that previously required an entire department. Someone else develops a product for a market too small to interest existing manufacturers.

A third person discovers that AI makes an entirely different business possible.

Most experiments won’t become successful companies.

But some will.

And as successful models emerge, larger organizations may consolidate them, develop supporting services, establish industry standards, and create entirely new commercial markets.

That’s how a collection of individual opportunities can eventually become a substantial industry.

There is an important qualification, however.

New businesses don’t necessarily create as many jobs as the businesses they replace. Some may generate considerable wealth while employing relatively few people.

And economic growth doesn’t guarantee that displaced workers will find comparable opportunities, particularly in the communities where they currently live.

Those are real problems that deserve serious attention.

But they are not reasons to assume that the number of jobs existing today represents the limit of future employment.

The Part Nobody Can Forecast

We’re investing extraordinary amounts of money in AI, robotics, data centers, and the infrastructure required to support them.

Much of the justification involves productivity improvements, cost reductions, and automating existing activities.

Those benefits may be substantial.

But I suspect the most important economic consequences won’t come simply from doing today’s work with fewer people.

They may come from doing things that aren’t economically possible today.

New products. New services. New markets. New ways of organizing businesses. New opportunities for individuals who previously lacked the resources to compete.

We cannot put reliable numbers on these possibilities because many of them haven’t been invented.

And that makes them easy to overlook in economic forecasts.

I’ve spent much of my professional life watching new technologies change existing businesses and create opportunities nobody anticipated.

Technology can make something possible. Someone still has to recognize that possibility, find customers, and turn it into a sustainable business.

So while everyone else is busy counting the jobs artificial intelligence may eliminate, I’m interested in the people discovering what they can do with it.

Somewhere, someone may already be building the next industry that nobody knows how to describe.

And twenty years from now, we may wonder how we ever managed without it.