Machines, Automation, and the Fear of Being Replaced
In 2023, a young woman in Bangalore lost her job.
She was a content writer for a digital marketing agency. She wrote product descriptions, blog posts, social media captions — the kind of writing that fills the internet. She was good at it. She understood tone, audience, the subtle art of making a shampoo sound like a life-changing experience.
One Monday morning, her manager called a meeting. The company had been testing an AI tool — a large language model that could produce content in seconds. Not as nuanced as hers, perhaps. Not as clever. But fast. And cheap. Astonishingly cheap. Where she cost the company forty thousand rupees a month, the AI cost a few hundred.
"We are restructuring," the manager said.
She was not alone. Her team of eight writers was reduced to two. The remaining two would "supervise" the AI's output — editing what the machine produced rather than creating from scratch.
She went home to her paying-guest accommodation in Koramangala, opened her laptop, and typed into the very tool that had replaced her: "How do I find a new career?"
The machine responded with cheerful, confident suggestions.
Look Around You
Think about your own work — or the work you are studying for. Could a machine do it? Not today, perhaps. But in ten years? Twenty?
Now think about the work your parents do. Could a machine do that?
This question used to be relevant only for factory workers and farmers. Today, it is relevant for everyone — writers, lawyers, accountants, radiologists, translators, customer service agents, even some doctors and engineers.
The fear of being replaced by a machine is older than you think. And the answer is more complicated than either the optimists or the pessimists will tell you.
The Oldest Fear in Economics
The fear that machines will destroy jobs is not new. It is as old as machines themselves.
In 1589, William Lee invented the stocking frame — a machine that could knit stockings far faster than a hand-knitter. He sought a patent from Queen Elizabeth I. She refused, saying: "Consider thou what the invention could do to my poor subjects who earn their bread by the knitting of stockings."
Elizabeth was not being sentimental. She was making an economic judgment: the social cost of unemployment outweighed the efficiency gains of the machine. (Lee eventually took his invention to France. Centuries later, Britain adopted it anyway.)
In the early 1800s, the Luddites — whom we met in an earlier chapter — smashed the power looms and shearing frames that were destroying their livelihoods. They were hanged for it.
In the 1930s, John Maynard Keynes coined the term "technological unemployment" — joblessness caused by the discovery of means to economize the use of labor "outrunning the pace at which we can find new uses for labour."
In the 1960s, President Lyndon Johnson's administration convened a commission on technology and automation, worried that computers would create mass unemployment.
Each time, the fear was real. And each time, the pessimists were mostly wrong — in the long run. Machines destroyed old jobs, but new jobs emerged. The economy adapted. Total employment grew.
But "mostly wrong in the long run" is not the same as "always wrong." And "the long run" can mean decades of suffering for the people whose jobs are destroyed right now.
Creative Destruction: The Theory
The Austrian economist Joseph Schumpeter, writing in the 1940s, described capitalism as a process of "creative destruction." Old industries die. New ones are born. The horse-drawn carriage gives way to the automobile. The telegraph gives way to the telephone. The typewriter gives way to the computer.
Each transition destroys jobs — coachmen, telegraph operators, typists — and creates new ones: auto mechanics, software engineers, social media managers. The process is painful for those caught in the destruction, but overall, it drives progress and raises living standards.
This story is true. But it is incomplete.
+--------------------------------------------------------------+
| WAVES OF AUTOMATION THROUGH HISTORY |
+--------------------------------------------------------------+
| |
| WAVE 1: MECHANICAL (1760s - 1850s) |
| Spinning jenny, power loom, steam engine |
| DESTROYED: Hand spinners, hand weavers |
| CREATED: Factory workers, mechanics, engineers |
| TIME TO ADJUST: ~60 years |
| |
| WAVE 2: ELECTRICAL (1880s - 1940s) |
| Electric motors, assembly lines, mass production |
| DESTROYED: Small craftsmen, artisan workshops |
| CREATED: Assembly line workers, electricians, managers |
| TIME TO ADJUST: ~40 years |
| |
| WAVE 3: DIGITAL (1960s - 2010s) |
| Computers, robotics, internet |
| DESTROYED: Typists, switchboard operators, some |
| factory workers |
| CREATED: Programmers, IT workers, web designers, |
| data analysts |
| TIME TO ADJUST: ~30 years |
| |
| WAVE 4: AI & INTELLIGENT AUTOMATION (2010s - ???) |
| Machine learning, large language models, robotics + AI |
| DESTROYING: Content writers, translators, data entry, |
| some legal work, some diagnostic medicine, |
| customer service, some coding |
| CREATING: ??? (This is the question) |
| TIME TO ADJUST: ??? |
| |
+--------------------------------------------------------------+
The ATM Paradox
One of the most cited examples in the "machines do not destroy jobs" argument is the ATM.
In the 1970s, automated teller machines began replacing bank tellers. The prediction was obvious: bank tellers would disappear. But they did not. The number of bank tellers in the United States actually increased from about 300,000 in 1970 to about 600,000 in 2010.
Why? Because ATMs made it cheaper to operate a bank branch. Banks opened more branches. And while each branch needed fewer tellers, the total number of branches grew enough to increase overall teller employment. The tellers' jobs also changed — from counting cash (which the ATM now did) to customer service, selling financial products, and building relationships.
This story is comforting. And it is true. But it has an important caveat: it worked because there was something else for the tellers to do. The ATM automated a task (dispensing cash), not a job (serving customers). As long as machines automate tasks rather than entire jobs, workers can shift to the remaining tasks.
But what happens when machines can do most or all of the tasks that make up a job?
The Loom Tells a Different Story
The ATM story has a less comforting counterpart: the power loom.
When the power loom was introduced in early 19th century England, it did not just automate a task within weaving. It automated weaving itself. A hand-loom weaver's entire set of skills — their knowledge of thread tension, their feel for the fabric, their years of practice — became irrelevant. A child minding a power loom could produce more cloth than a master weaver.
The hand-loom weavers did not transition into new jobs within the textile industry. They were simply destroyed as an occupational class. Many starved. Some rioted. Some emigrated. Their skills, accumulated over generations, became worthless overnight.
Eventually, the economy did create new jobs — but not for the weavers. The new jobs required different skills, in different places, for different people. The weavers of 1810 did not become the factory workers of 1840. Their children did, perhaps. Or their grandchildren. "Creative destruction" creates in the long run, but it destroys right now — and the people who are destroyed do not always live to see the creation.
"In the long run we are all dead." — John Maynard Keynes
Why This Time Might Be Different
Every generation has said "this time is different" about automation, and every generation has been wrong. So we should be humble about predictions.
But there are reasons to think the current wave of automation — driven by artificial intelligence — poses challenges that previous waves did not.
Previous automation replaced muscle. AI replaces mind.
The first three waves of automation affected primarily physical labor. Machines lifted, carried, assembled, and manufactured better than human bodies could. But human minds remained essential — for judgment, creativity, communication, complex decision-making.
AI changes this. A large language model can write, translate, summarize, analyze, code, and reason. A machine vision system can diagnose diseases from medical images. An AI system can draft legal contracts, generate architectural plans, compose music, and create art.
For the first time in history, machines are entering the domain of cognitive work — the very work that people were told to pursue because it was "safe from automation."
The pace is faster. The Industrial Revolution took decades to transform the economy. The AI revolution is moving in years. A technology that barely existed in 2020 is transforming industries by 2025. Workers, institutions, and education systems cannot adapt at this speed.
The scale is global. When British factories destroyed Indian weavers, the destruction was limited by geography and trade routes. AI is available everywhere, instantly. A call center in Gurgaon can be replaced by an AI chatbot on a server in Oregon overnight. There is no time cushion, no geographic buffer.
But there is a counter-argument.
AI is good at tasks, not at being human. An AI can write a blog post, but it cannot sit with a grieving family. It can diagnose a disease from an X-ray, but it cannot hold a patient's hand. It can generate legal text, but it cannot understand what justice feels like to a person who has been wronged.
The jobs that are hardest to automate are those that require physical dexterity in unpredictable environments (plumbing, electrical repair, gardening), emotional intelligence (nursing, teaching, social work), and genuine creativity (the kind of art that surprises even the artist).
The irony is bitter: the jobs that are safest from AI are often the jobs that are lowest paid. The plumber is safer than the programmer. The nurse is safer than the analyst. The economy may be turned upside down — the work of the hand becoming more valuable than the work of the mind.
What Actually Happened
In 2024, IBM announced that it would pause hiring for roles that could be performed by AI, affecting roughly 7,800 positions — mostly in back-office functions like HR. Numerous tech companies followed with layoffs partly attributed to AI-driven efficiency gains.
But in the same year, new roles emerged: AI prompt engineers, AI trainers, AI ethics specialists, AI integration consultants. Companies hired people to manage, supervise, and correct AI systems.
The pattern was familiar — old jobs dying, new ones being born. But there was a catch: the new jobs required different skills, in different locations, at different pay scales. An HR assistant in Pune who lost her job could not simply become an AI prompt engineer in San Francisco. The creation was real, but it did not help the people who experienced the destruction.
The Race Between Education and Technology
The economists Claudia Goldin and Lawrence Katz, in their book The Race Between Education and Technology, made a powerful argument: throughout the twentieth century, the key determinant of whether workers benefited from technological change was education.
When education advanced faster than technology — as it did in the early and mid-twentieth century, when high school and college education expanded massively — workers were equipped for new jobs and wages rose broadly.
When technology advanced faster than education — as it has since the 1980s — the benefits of growth went disproportionately to those with the highest skills, while those without fell behind.
The lesson: the impact of automation is not determined by the machines alone. It is determined by how well we prepare people to work alongside — or instead of — the machines.
In India, this challenge is acute. The education system produces millions of graduates every year, but many lack the skills that the evolving economy demands. A 2019 report by the Azim Premji Foundation found that nearly half of India's engineering graduates were unemployable by industry standards. If AI transforms the economy faster than India can transform its education system, the result will be mass unemployment among the very people who were told that education was their ticket to prosperity.
"The illiterate of the twenty-first century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn." — Alvin Toffler
What Happens When Machines Can Do Everything?
Let us push the thought experiment to its extreme. What happens if machines can eventually do everything humans can do — physical and cognitive — but cheaper and better?
This is no longer pure science fiction. It is a scenario that serious economists and technologists are discussing.
If it happens — and "if" is important — the result would be paradoxical. Enormous wealth would be created. Goods and services would become almost free to produce. But the traditional mechanism for distributing that wealth — paying people wages for work — would break down. If machines do the work, who earns the money to buy the products the machines make?
This is where the debate over Universal Basic Income (UBI) enters.
The idea is simple: every citizen receives a regular cash payment from the government, enough to meet basic needs, regardless of whether they work. No conditions, no means testing, no bureaucratic gatekeeping.
Proponents argue that UBI would provide a floor of security in an age of automation. It would give people the freedom to retrain, start businesses, care for family members, or pursue creative work. It would replace a tangle of welfare programs with a single, efficient transfer.
Critics argue that UBI would be ruinously expensive, that it would discourage work, and that it would drain resources from targeted programs that help those who need it most.
India has experimented with something resembling UBI. The PM- KISAN scheme provides Rs. 6,000 per year to small and marginal farmers — a tiny amount, but a direct cash transfer to over 100 million families. The Telangana state government's Rythu Bandhu scheme goes further, providing investment support to farmers based on land holdings.
These are not UBI in the pure sense — they are targeted and conditional. But they represent a shift in thinking: from providing work (as MGNREGA does) to providing income directly.
The debate is far from settled. But the question it raises is fundamental: in a world where machines can produce everything, how do we distribute the wealth they create?
+--------------------------------------------------------------+
| THE AUTOMATION QUESTION |
+--------------------------------------------------------------+
| |
| IF MACHINES CAN DO MORE AND MORE... |
| |
| OPTIMISTIC VIEW | PESSIMISTIC VIEW |
| ============================|==============================|
| New jobs will emerge | Not enough new jobs |
| (they always have) | (this time is different) |
| | |
| Humans will do what | Machines will learn to do |
| machines cannot | that too |
| | |
| Education will adapt | Education cannot keep up |
| | |
| Productivity gains will | Gains will go to capital |
| benefit everyone | owners, not workers |
| | |
| POLICY IMPLICATIONS: | POLICY IMPLICATIONS: |
| - Invest in education | - Universal Basic Income |
| - Retrain displaced workers | - Wealth taxes |
| - Trust the market | - Shorter work week |
| | - Public ownership of AI |
| | |
| TRUTH: Probably somewhere in the middle. |
| The outcome depends on CHOICES — political, institutional, |
| and social — not on technology alone. |
| |
+--------------------------------------------------------------+
Automation in India: A Special Challenge
India faces the automation challenge from a unique position.
Unlike the United States or Europe, where automation threatens existing middle-class jobs, India faces automation before it has fully created those jobs. India has not yet completed its industrial transition. Hundreds of millions of workers still depend on agriculture and informal labor. The manufacturing sector — which was supposed to absorb these workers, as it did in China and Korea — is being automated before it has fully grown.
This is the "premature deindustrialization" that economist Dani Rodrik has warned about. Countries like India may find that the manufacturing pathway to prosperity — the path every previously successful economy has followed — is closing. Robots can now do what cheap labor once did. Why would a global company build a factory in India with 10,000 workers when it can build a factory in Germany with 1,000 robots?
The implications are staggering. If India cannot absorb its massive young population into productive employment — and India adds roughly 12 million people to the working-age population every year — the result could be a permanent class of underemployed young people. The demographic dividend that economists have celebrated could become a demographic disaster.
This is not inevitable. India has strengths — a large domestic market, a growing services sector, a young and increasingly educated population. But it requires policy choices that are clear-eyed about the challenge.
The Human Question
Behind all the economic analysis, there is a deeper question.
Work is not just a way to earn money. It is a source of meaning, identity, and dignity. When you meet someone new, one of the first questions you ask is: "What do you do?" The answer places them in the social world. It tells you — and them — who they are.
What happens when machines take that away?
A factory worker who loses their job to a robot does not just lose income. They lose their place in the world. They lose the daily rhythm that structured their life. They lose the camaraderie of coworkers. They lose the identity that came from being good at something.
Studies of communities where factories have closed — the American Rust Belt, the British Midlands, the mill towns of Maharashtra — consistently find the same pattern: not just economic decline, but social disintegration. Alcoholism rises. Families break. Civic institutions wither. People lose hope.
This is not something that a UBI check can fix. Money helps. But meaning — the sense that your work matters, that you are contributing something, that you are needed — is harder to replace.
The challenge of automation is ultimately a challenge of meaning. If machines can do everything we can do, what is the point of us?
The answer, perhaps, is that we are not meant to be useful. We are meant to be alive — to love, to create, to wonder, to care for each other in ways that no machine can replicate. But building an economy and a society around this insight — rather than around the equation of human value with economic productivity — would require a revolution in thinking that we have barely begun.
Think About It
Name three jobs that existed fifty years ago and no longer exist. Name three jobs that exist today and did not exist fifty years ago. What pattern do you see?
If AI could do your job (or the job you are studying for) better and cheaper, what would you do? How would you feel?
"Machines create more jobs than they destroy." This has been true historically. Do you think it will remain true? What would it take for it to stop being true?
Should the government provide a Universal Basic Income to all citizens? What would be the benefits? What would be the costs? Who would pay for it?
If machines could do all the work, and humans did not need to work to survive, would that be paradise or a nightmare? Why?
The Bigger Picture
Every wave of automation has created fear. And every time, so far, the fear has been partly wrong. New jobs have emerged. The economy has adapted. Living standards have risen.
But every time, the fear has also been partly right. Real people have suffered. Real communities have been destroyed. The "adjustment" that economists describe in a sentence can take a generation to complete, and the generation that lives through it pays a price that no subsequent prosperity can fully repay.
The question is not whether technology will advance. It will. The question is who will benefit and who will bear the costs.
If we leave the answer to the market alone, the historical pattern is clear: the owners of capital — who own the machines — will capture most of the gains, and the workers — who are replaced by the machines — will bear most of the costs.
If we intervene — through education, through social protection, through taxation that shares the gains widely, through policies that ensure technology serves people rather than the other way around — the outcome can be different.
Technology is not destiny. It is a tool. And like all tools, its impact depends on who holds it and for what purpose.
The young writer in Bangalore who lost her job to an AI did not lose because of technology. She lost because the economy in which she lived valued cost reduction over human creativity, efficiency over dignity, and shareholder returns over workers' livelihoods.
Those are not technological choices. They are human ones. And human choices can be changed.
"We are called to be architects of the future, not its victims." — R. Buckminster Fuller
The machines are coming. The question is not how to stop them. The question is how to ensure that the future they build has room for all of us.
In the next chapter, we follow a different kind of movement — not machines moving into human work, but human beings moving to where the work is. The story of migration: why people leave home, what they find, and what they leave behind.