The automation-anxiety tax
The fear of artificial intelligence is now the median mood at work. Long before any job is automated, that fear is quietly collecting a levy on the ambition and nerve most careers run on.
Consider a competent person in the middle of a good career. Call the details ordinary, because they are. She is forty-one, three promotions in, respected on her team, and last Tuesday she did something she would not have done two years ago. A stretch assignment came up, the kind she would once have raised her hand for without thinking, and she let the moment pass. Someone younger took it. That night, instead of sleeping, she opened a browser tab and worked through a tutorial for a tool she does not enjoy and does not need yet, because a headline that afternoon had used the word obsolete and it lodged somewhere she could not reach.
Nothing in that day would ever show up in a layoff statistic. No job was lost. No task was automated. And yet something was quietly spent. She made herself smaller and called it prudence. She learned something out of dread rather than interest and called it staying current. Multiply that Tuesday across millions of desks and you get a cost no productivity report will ever capture, because it is paid in the things people stop doing.
This is the automation-anxiety tax. It is the levy the fear of artificial intelligence collects on ambition and nerve, and the striking thing about it is that it comes due long before, and often instead of, any actual automation. The machine does not have to take your job to change how you behave inside it. The anticipation is enough.
A fear that is now the baseline
This is no longer a fringe worry held by the technophobic. It is the median position. When the Pew Research Center surveyed 5,273 employed U.S. adults in late 2024, 52 percent said they were more worried than hopeful about the future use of AI in the workplace, far outnumbering those who felt hopeful. Nearly a third, 32 percent, expected AI to lead to fewer job opportunities for them personally over the long run. These are not the conclusions of people who have read the displacement research. They are a mood, and the mood is caution.
Ask people to predict their own fate and the number gets smaller but hits harder. In a February 2026 poll of 23,717 employed adults, Gallup found that 18 percent thought it very or somewhat likely their job would be eliminated within five years because of AI or automation. Among people who already work at organizations that have adopted AI, that share rose to 23 percent. Proximity to the technology does not soothe the fear. It sharpens it. The people closest to the tools are the ones bracing hardest.
The tax is paid in caution
A fear that widespread does not sit still inside people. It leaks into decisions. The clearest place to watch it is in how willing workers are to move. For most of the last decade, changing jobs was how ambitious people got paid and got ahead. Then it stopped. According to the Bureau of Labor Statistics, the rate at which Americans quit their jobs has fallen to around 1.8 percent, its lowest reading outside the pandemic since 2014, down from a peak near 3 percent during the resignations of 2022.
Some of that is a cooler labor market. But part of it is a posture. When the ground feels like it might shift, you plant your feet. We have written about the quiet mass decision to stay put and about what the collapse in job-hopping is actually telling us, and the automation story pours directly into that reservoir of caution. You do not leave a job you can tolerate for one you might love if some part of you believes the whole category could be hollowed out by a model release you cannot predict. Ambition requires a bet on the future. The tax is what makes the future feel unbettable.
Reskilling as tribute
There is a second way the tax gets collected, and it looks, deceptively, like initiative. The World Economic Forum's Future of Jobs report, drawing on employers who together account for millions of workers, projects that 39 percent of the skills workers have today will be transformed or made outdated between 2025 and 2030, and that 40 percent of employers expect to reduce their workforce where AI can automate the task. Read against a frightened workforce, those numbers stop being a forecast and become a standing instruction: keep running or be left behind.
So people run. They learn tools at eleven at night. The trouble is the motive. Learning driven by curiosity compounds into mastery. Learning driven by dread compounds into exhaustion, because the goalposts are built to move and the runner knows it. It is the difference between training for a race and running from something, and only one of those makes you faster over time. We have written before about how quickly a skill now loses its value. The anxiety tax is what turns that fact from a plan into a panic, and it quietly bills the person twice: once for the learning, and again for the nagging certainty that the learning will not be enough.
The machine does not have to take your job to change how you behave inside it. The anticipation is enough.
What the fear crowds out
Here is what the tax actually costs, and it is not the thing the headlines warn about. The danger is not mainly that AI will take a specific person's specific job. For most people, on most days, it will not, at least not soon. The danger is quieter and more certain than that. Fear converts a useful question into a useless one.
The useful question, the one a mid-career professional is quietly built to ask, is a question of fit: does this particular job, in this particular place, still match who I have become and what I now need from my work. That question is generative. It leads somewhere. It sits underneath every good move anyone has ever made, and it is the same question underneath why so much workplace distress turns out to be a mismatch rather than a personal failing. But you can only ask it from some position of security. Frightened people do not ask whether a seat fits them. They ask only whether they can keep it. Survival questions have exactly one answer, which is to hold on to whatever you already have, tighter.
That is the tax in its final form. It does not empty your bank account, and it may never touch your job. It empties your agency. It takes a person who should be asking where they want to go and reduces them to bracing for what might be done to them. The automation may or may not arrive on the schedule the fear insists on. The bracing is already here, and the bracing is the cost.
None of this argues that the technology is harmless or the disruption imagined. The models are real and some of the displacement will be too. It is an argument about where to spend your attention while the future is still unwritten, which is to say now. A machine that does not yet exist cannot make your choices for you. Fear can, if you let it hold the pen. The most expensive thing the anxiety takes is not the job. It is the version of you that was still willing to want something and go get it.