0
0
You are at your desk. Someone drops a task into a chat box — the kind of task you spent years learning to do properly — and nine seconds later it comes back done. Not perfectly. But done well enough that nobody in the room says anything.
Everyone goes back to work. You go back to work. And somewhere in that quiet afternoon a question arrives that you would never say out loud: will AI take my job?
That is the question people type into a search bar. It is not the question that keeps them awake. Underneath it there is a smaller, sharper one: if a machine can do the thing I am good at, am I still worth anything?
This is about both. The first question has some surprisingly honest answers. The second one has an answer much older than any machine.
Will AI Take My Job? What the Famous 47% Actually Measured
You have probably met the number. Nearly half of all jobs at risk from automation. It turns up in headlines, in meetings, in nervous jokes over lunch.
It comes from a 2013 paper by two Oxford researchers, Carl Benedikt Frey and Michael Osborne, called The Future of Employment. They scored 702 occupations for how easily a computer could handle the tasks inside them. About 47% of total US employment landed in their “high risk” group.
Here is the part that almost never gets repeated. In the same paper, the authors wrote this:
“We make no attempt to estimate the number of jobs that will actually be automated, and focus on potential job automatability over some unspecified number of years.”
Read it again slowly. They were not predicting job losses. They were measuring what is technically possible, on a timeline they openly refused to pin down — “perhaps over the next decade or two,” as they put it a few lines later.
So the most quoted number in the whole conversation about work and machines is not a forecast at all. It is a map of what machines could potentially reach, published with its own warning label attached. The warning label fell off somewhere between the paper and the headline.
That does not mean nothing is happening. It means the number you have been carrying around in your chest was never measuring the thing you thought it was measuring.
The Cash Machine Story That Nobody Finishes
There is a story people tell about cash machines, and they almost always stop halfway through it.
The first half is familiar. Cash machines arrived, counting money stopped being a human job, and bank tellers were finished. The economist James Bessen went and looked at what actually happened, and published it in the IMF’s Finance & Development magazine in March 2015.
Between 1988 and 2004, the number of tellers needed to run an average urban branch fell from 20 to 13. The machines really did take that work. That part of the story is true.
But a branch that runs on 13 people instead of 20 is a cheaper branch. So banks opened more of them. Urban branches rose by about 43%. More branches, fewer people in each one. Bessen’s finding was that the total number of teller jobs did not go down.
And the job itself changed shape. Once a machine handled the counting, what a teller was for became the part no machine could do — knowing a customer, sorting out a problem, noticing when something was wrong.
Be careful with this story, though. It is one job, in one industry, over one stretch of years, in one country. It is not a promise. Plenty of jobs have vanished and never come back. The honest version is narrower and more useful: when a machine takes the measurable part of a job, the unmeasurable part does not automatically become worthless. Sometimes it quietly becomes the whole job.
Why It Hurts Even When Your Job Is Fine
Here is something that should not make sense. Many of the people carrying this fear are not about to lose anything. The pay still arrives. The work still comes in. The dread sits there anyway.
Two economists, Liliana and Rainer Winkelmann, went looking for why losing work hurts people so much. They followed working-age men in Germany over years rather than asking them once, and published the result in the journal Economica in 1998. Unemployment had a large negative effect on how satisfied people were with their lives. No surprise there.
The surprise was in the breakdown. The damage caused by losing the income was the smaller part. In their words, “the non-pecuniary effect is much larger than the effect that stems from the associated loss of income.”
Most of the pain of losing work is not about the money.
Which tells you what this fear is actually made of. If money were the whole story, a fully paid worker watching a machine do their task would feel nothing at all. But they do feel something, because what is being tested in that moment is not their income. It is whether they are needed.
You can see the same thing in how people are behaving at work right now. We looked at job hugging — 57% of workers saying they stay where they are out of fear rather than loyalty. Fear like that is not a salary calculation. It is a worth calculation.
The Sum Nobody Actually Lives By
Follow the fear all the way down and you find a very simple sum sitting at the bottom of it.
If my value is what I produce, then anything that produces faster makes me worth less.
That sum is tidy. It is also something almost nobody believes the moment they look at an actual person. You do not love a child for their output. You do not sit with a sick friend because of their productivity. Nobody has ever stood up at a funeral and talked about units completed. The second it is someone we love, we drop the formula instantly — and we never notice we dropped it.
Which raises a fair question. If we refuse to use the output formula on anyone we care about, why do we keep using it on ourselves?
The oldest account anyone has of where human beings come from does not open with a job. It opens with worth already settled: a person described as made on purpose, carrying something of God in them, and only after that is any work handed over. The order is the whole argument. The dignity came first. The task came second, and it was given to someone who already had it.
You do not have to believe a word of that to notice what it would mean if it were true. It would mean your worth was never yours to earn in the first place — which also means it was never a machine’s to take. (If you want that argument at length and in formal language, the Vatican took up the same question in a recent document on artificial intelligence, and we went through it here.)
What Was Actually Being Tested
The tools are fast. They are going to keep getting faster, and nothing here is going to talk you out of noticing that. Learn what they can do. Learn what they cannot. Move towards the part of your work a machine has no access to, because that part is becoming more valuable, not less.
But the private test you were running that quiet afternoon — the one where you checked whether you were still worth anything — was never a real test. The result was in long before you started it. It was in before you learned the skill, before you got the job, before you were of any use to anybody at all.
If that question has been sitting with you for a while, it is worth looking at in daylight rather than at two in the morning. The free What Is My Purpose? assessment takes a few minutes and asks a different question than the one you have been asking — not what you are good at, but what you are for. That turns out to be the one that was actually bothering you.
The machine does your task. It has never once wondered whether it matters. You have. That is not a small difference. It might be the entire difference.
Questions People Ask About AI and Work
Will AI take my job?
Nobody can answer that for one specific job with confidence, and anyone who sounds certain is guessing. What can be said is that the famous estimate behind the fear — a 2013 Oxford paper by Carl Benedikt Frey and Michael Osborne, which scored 702 occupations and placed about 47% of US employment in a “high risk” group — was explicitly not a prediction of job losses. The authors wrote that they made “no attempt to estimate the number of jobs that will actually be automated.” It measured technical possibility, not outcomes. The more useful question for any individual job is which parts of it are measurable and repeatable, and which parts depend on judgment, trust and dealing with people.
Is it true that 47% of jobs will be lost to automation?
No, and that is not what the research claimed. The 47% figure comes from the 2013 paper “The Future of Employment” by Carl Benedikt Frey and Michael Osborne of Oxford. They estimated that about 47% of total US employment fell into a “high risk” category for computerisation “over some unspecified number of years, perhaps a decade or two.” They also stated plainly that they made no attempt to estimate how many jobs would actually be automated. The number describes what machines could potentially do, not a measured or predicted level of job loss.
Did bank tellers disappear after cash machines arrived?
No. Economist James Bessen, writing in the IMF’s Finance & Development magazine in March 2015, found that between 1988 and 2004 the number of tellers needed to run an average urban bank branch fell from 20 to 13 — but because branches became cheaper to operate, the number of urban branches rose by about 43%, and the total number of teller jobs did not decrease. The role shifted towards customer relationships, the part of the work the machines could not do. This is one occupation, in one industry, over one period, and it is not a general rule about automation.
Why does the idea of being replaced by AI feel so personal?
Because the pain of losing work is mostly not financial. Economists Liliana and Rainer Winkelmann, studying working-age men in Germany over several years and publishing in the journal Economica in 1998, found that unemployment had a large negative effect on life satisfaction, and that “the non-pecuniary effect is much larger than the effect that stems from the associated loss of income.” What work provides beyond money — being needed, being competent, being counted on — accounts for most of what hurts when work goes. That is why watching a machine do a task well can feel threatening even to someone whose pay is not at risk at all.
How do I stop tying my self-worth to my job?
Start by noticing that almost nobody applies the output formula to other people. Few would say a child, a sick friend or an elderly parent is worth less because they produce less — yet many apply exactly that standard to themselves. The inconsistency is the clue: worth-as-output is not a belief most people actually hold, it is a habit of self-assessment. Three steps that cost nothing: name one relationship where you are valued for being there rather than for performing; spend a little time each week on something you are deliberately bad at; and separate the sentence “my work is needed less” from the sentence “I am needed less,” because they are not the same sentence and only one of them is about you.
What Do You Think?
Here is the part worth arguing about. When a machine can do a task better than a person, does the person’s work become less valuable — or does the part only a person can do become more valuable? There is a real case on both sides. Tell us which way you lean in the comments, and if you have watched this happen in your own line of work, say what actually changed.
Share This
- The famous “47% of jobs at risk from AI” number was never a prediction. The researchers wrote that they made “no attempt to estimate the number of jobs that will actually be automated.” The warning label fell off somewhere between the paper and the headline. https://bgodinspired.com/index.php/personal-growth-and-life-skills/will-ai-take-my-job/
- Cash machines cut the tellers needed per bank branch from 20 to 13. Then branches got cheaper, so banks opened 43% more of them — and teller jobs did not go down. The machine took the counting. The job became the part it could not do. https://bgodinspired.com/index.php/personal-growth-and-life-skills/will-ai-take-my-job/
- Economists found that most of the pain of losing work is not about the lost income. Which explains why watching a machine do your job can hurt even when your pay is perfectly safe. It was never the money being tested. https://bgodinspired.com/index.php/personal-growth-and-life-skills/will-ai-take-my-job/
