This column was originally published on the Ball State University Center for Business and Economic Research Weekly Commentary blog.
By Michael J. Hicks
May 3, 2026
I’ve spent a lot of time studying the potential labor market effects of artificial intelligence. I
think the greatest uncertainty isn’t what AI will do to workers, but what it will do to
communities.
There is an enormous amount of fearmongering about AI. It is easy to find a smartly written
piece warning of machines on the loose and the end of humanity. But, it is worth remembering
that alarmism about the labor market effects of the power loom caused actual riots in Britain two
centuries ago, leading to trials and executions.
Technology doesn’t replace jobs. It replaces tasks that people perform. If you suppose that
technology kills jobs, you also need to admit that the wheel was probably the biggest job killer in
history. Somehow, we muddled along.
Technology creates new demand in two ways. First, a widely adopted technology reduces costs
for something, say, the transport of goods or the manufacture of clothing. Second, technology
frees up spending power to buy other things, like leisure or health care.
This is a fancy way of saying technology makes us richer.
Labor economists looking at rapid technological change take these things into account. In doing
so, there are few, if any, examples of job-killing technology. Take the most recent example: the
vast productivity growth and imports of manufacturing goods in the U.S.
From 1979, when factory employment peaked, until the last month of the Great Recession, we
lost 7.9 million factory jobs. But, over the same period, we gained 68.4 million jobs. So, for
every factory job we lost, we gained 8.7 non-factory jobs, and inflation-adjusted disposable
income more than doubled.
One cannot objectively examine national labor markets from 1979 to 2009 and conclude that it
was anything but a smooth and advantageous adaptation to technological change. I’d wager we’ll
look back on AI in 40 years and render the same judgment.
Still, few Americans living in the Midwest feel that it was an easy transition. Those living in the
South, Northeast or West Coast probably wonder what the fuss is about.
What’s missing from the decline of manufacturing employment is precisely what we are missing
in the discussion about artificial intelligence. The real shocks, and the pain that accompanied
those shocks, were isolated.
For AI, the local effects have two differing components.
If AI makes us richer, as it surely will, it will unleash demand for tasks, and jobs, that may be
hard to imagine now. This is certain to be the case. I can foresee a wide variety of new
occupations that use AI in wholly different ways. I suspect most people can think of a few things
they’d pay to do, or pay not to do, that AI can help with.
If I can think of these tasks, I’m sure there are hundreds of thousands of entrepreneurs thinking
about, and implementing, these tasks.
But tasks that use AI, like the tasks of every other technological improvement before it, from the
wheel to the internet, will require more education and training. What is most interesting about
the AI revolution is that the education requirements appear to be more like what we expected of
education one century, or even five centuries, ago than what we think about today.
AI solves a lot of routine, cognitive tasks, including software coding and interpreting detailed
statistical modeling quickly and effectively. This is the type of work I spent countless hours in
graduate school, and years thereafter, mastering. So, I’m not romanticizing my years in college.
What AI cannot do well, and is unlikely to do well, are the more subtle human cognitive
functions. AI can aid in the construction and interpretation of vast finance models, the weather
and interplanetary travel. AI is very bad at interpreting a Jane Austen book, our preferences
about weather or the benefits and costs of interplanetary travel. This serves to reduce the cost of
the former and increase the costs of the latter skills.
AI is probably the most important, and beneficial, technology for the humanities since the
printing press. If you want to ensure your kids are ready for AI, introduce them to Jane Austen,
Marcus Aurelius, Shakespeare and Hemingway. Those lessons are hard for AI. They are hard for
students, too, but the lessons are complex and subtle, and enduring.
Strong abstract thinking skills are what the AI age requires. That education is very expensive, in
both time and treasure. An education for the AI world will likely look more like a traditional
liberal education, with a healthy dose of problem solving, than what now passes for the career-
ready focus.
These are the demand-side aspects of AI; there are also supply-side effects.
Most data centers are looking for cheap land and water (and cheap local government). About
70% of data centers in the U.S., and much more of the square footage, do AI processing for users
outside the local area. These data centers need almost no workers but consume a great deal of
power and water.
The remainder of data centers, the smaller 30% of facilities, are placed around the skilled
workers who need proximity to high-end computing capacity. Think financial services, health
care diagnostics, national security applications and other sensitive computing processes.
AI will also require much more electric power generation. Most of this will be wind and solar,
but a growing share will be nuclear. While those positions create jobs, most are footloose, so we
will see more jobs in low-cost, low-wage areas building solar panels and turbines.
This world will create two different outcomes for communities hosting data centers. One is land-
and capital-intensive, without much need for people. The other needs high-end labor, that is,
workers who have the four to eight years of post-secondary education it takes to complement the
benefits of AI.
The power generation will contribute to large regional differences as well. Both the demand- and
supply-side effects of growing AI have disparate effects on individuals. But, as with earlier
technologies, the most visible differences won’t be between people, but places.
That will be the bigger challenge of the AI world we are entering.
Michael J. Hicks is professor of economics and the director of the Center for Business and Economic Research at Ball State University. He previously served on the faculty of the Air Force Institute of Technology’s Graduate School of Engineering and Management and at research centers at Marshall University and the University of Tennessee. His research interest is in state and local public finance and the effect of public policy on the location, composition, and size of economic activity.
The views expressed here are solely those of the author, and do not represent those of funders, associations, any entity of Ball State University, or its governing body. Also, the views and opinions expressed do not necessarily reflect the views of The Indiana Citizen or any other affiliated organization.
