By Dr. James Stapleton

CEO, Codefi Foundation on Rural Innovation

In every room where I present this work — hospital boardrooms, chamber luncheons, community colleges — the same question eventually stands up. Sometimes it is asked carefully, sometimes bluntly, but it is always in the room: if these tools are as capable as you say, is this how our jobs disappear?

The question deserves respect, because the anxiety behind it is real and widely shared. Pew Research Center finds that about half of American workers — 52 percent — are worried about how AI will be used in the workplace, roughly one and a half times the share who feel hopeful. Workers are not being irrational. They are applying a lesson history actually taught: that when a powerful technology meets a workforce, some of that workforce gets displaced. What I want to show in this post is that the lesson carries a hidden premise — and in our economy, the premise is false.

The displacement story assumes workers to spare

Every displacement narrative — from the mechanized loom to the assembly robot — runs on the same background condition: a surplus of labor relative to the work. Automation “takes” a job when there is someone left holding nothing after the machine is installed. The fear borrows its logic from economies where workers outnumbered the work.

Now recall what the first post in this series established about the economy we actually have. The workers of the 2030s have already been born, and there are not enough of them; the U.S. fertility rate has run below replacement since 2007. Missouri has roughly 87 available workers for every 100 open jobs before the rural multiplier is applied. Federal projections show labor force participation falling through 2034. We are not an economy with workers to spare. We are an economy where the work outnumbers the workers — and the gap is structural.

Run the displacement fear through that reality and it inverts. You cannot be replaced by AI when there is no one to replace you with. In a labor-short economy, AI does not walk in and take a chair from a worker. It fills the chair that has been sitting empty — the posting that never filled, the position absorbed “for now” by everyone else, the capacity gap a vacancy already created.

What has actually happened: three years of evidence

This is no longer only a logical argument; it is now an empirical record. Researchers at Yale's Budget Lab, working with the Brookings Institution, have tracked the U.S. labor market month by month since ChatGPT's launch — the fastest technology adoption ever measured, as the last post showed. Their finding, across nearly three years of data: no discernible disruption. The occupational mix has remained strikingly stable, changing no faster than it did before AI arrived. Employment and wages in AI-exposed occupations are statistically indistinguishable from everyone else's. And when they examined unemployed workers specifically — the place displaced people would have to show up — they found no rising share of AI-exposed occupations among them, regardless of how long people had been unemployed.

Honesty requires the caveats, and they matter. The researchers themselves say “for now” and monitor monthly; a three-year window is early; and there is credible evidence that entry-level hiring in the most exposed occupations has softened for recent graduates — a real signal about how career on-ramps are shifting, and one more reason training pathways that do not depend on the traditional entry ladder matter. But the headline conclusion stands: the predicted apocalypse is not in the data. Stability is.

Meanwhile, the demand side is doing the opposite of shrinking

Displacement also requires demand for the work to hold still while machines absorb it. In the industries where Missourians actually work, demand is doing the opposite — driven by the same demographics that created the shortage. The Census Bureau reports that from 2020 to 2024, the population 65 and older grew 13 percent while the working-age population grew just 1.4 percent — and that older adults now outnumber children in nearly half of U.S. counties, concentrated in exactly the kind of sparsely populated places that make up rural Missouri. The people who need the most care are multiplying fastest precisely where the workforce is thinnest.

Project that forward and the vise closes harder. The Bureau of Labor Statistics expects healthcare and social assistance to add nearly two million jobs by 2034 — roughly two-fifths of all projected U.S. employment growth, expanding almost three times as fast as the overall economy — driven by aging and chronic disease. And as Missouri's own hospital data showed in the first post, the hardest positions to fill are already the ones requiring the most training. There is no version of this arithmetic in which Missouri healthcare — or manufacturing, or agriculture, or public service — runs out of work for trained people. There are only versions in which the work goes undone.

The honest risk — and the hedge that wins either way

None of this means nothing changes. Tasks are already shifting; jobs are being recomposed around the tools; the on-ramps into some careers are moving. So let me state the real risk plainly, because workers deserve the truth rather than reassurance: in a labor-short economy, the risk is not that your job disappears. It is that capability decides who advances — and, as the last post showed, the benefits of these tools are currently concentrating among the already-trained. The question facing a Missouri worker is not “will AI take my job?” It is “will I be the person in my department who can direct it?”

Training is the hedge that wins in every scenario. If AI remains what the evidence says it is today — an augmenting technology in a labor-starved economy — the trained worker is simply more productive, more valuable, and harder to lose. If the technology someday becomes more disruptive than the current data shows, the trained worker is the one who directs the transition instead of absorbing it. There is no future in which knowing how these tools work leaves a worker worse off. The worker who learns to direct them does not become replaceable. She becomes the person her department cannot function without.

What to do with this

If the question that opened this post is the one that has kept you — or your team — from starting, consider it answered and start. Codefi's training begins with AI literacy, including what these tools cannot do and how to use them safely in regulated work, and builds to role-specific capability and credentials earned on the job, with grant-funded scholarship access available to eligible Missourians: codefiworks.com/ai-training.

With the fear addressed, the series turns from defense to offense. The next post is for the people responsible for regions rather than jobs — economic developers, chambers, civic leaders — because the same argument that protects a worker's future is the argument that decides where investment lands. Readiness is about to start choosing winners among regions. Missouri should be one of them.

Learn fast. Build right. Gain traction.