“All make-up is a lie.” – Morrissey
“Tech will destroy more jobs than it creates.” – Vinod Khosla
“Aasman pe hai khuda aur zameen pe hum, aaj kal wo is taraf dekhata hai kam.”- Sahir Ludhianvi
The tech bros are making money. The PR agencies, the consultant and the thinktank ecosystem that keeps the mahaul around artificial intelligence (AI) going, is making money. The chief executives who are doubling AI spending despite 95 percent of organisations getting zero return are, for now, keeping their jobs.
It is always safer to fail conventionally than to succeed unconventionally.
But what about the 23-year-old data analyst, the copywriter, the call centre agent, the junior software engineer? They are being told to reskill. Into what exactly, is left as an exercise for them to figure out.
The simplistic answer to the question whether AI will destroy jobs is yes because that’s what increasing productivity means – fewer people are needed to carry out an economically value adding activity. But the actual answer is quite complicated.
So, while technology destroys jobs, it also creates new jobs and new industries. Consider the horse, once indispensable to transportation, its economic value was upended by the arrival of the railways and the automobile.
With that shift, a range of livelihoods gradually disappeared – intercity riders, carriage drivers within towns, farms that bred horses, workshops that built carriages and even those employed to clear horse manure from city streets.
Yet technological change did not simply eliminate work; it reshaped it. New jobs emerged in factories manufacturing railway coaches and automobiles. Drivers and locomotive operators were needed to run the new machines. The construction of railway lines and, later, national highways generated employment on a massive scale.
Technological breakthroughs also tend to trigger further innovations. As Martin Wolf notes in The Crisis of Democratic Capitalism: “Electricity brought refrigeration, the telephone, the skyscraper, air-conditioning and the early computer.” Each wave of innovation sets off another – creating fresh industries and new forms of work in the process.
Take the case of the mass production of cars which started happening in the early twentieth century. Nobel Prize winning economists Daron Acemoglu and Simon Johnson point out in Power and Progress – Our Thousand Year Struggle Over Technology and Prosperity that mass car production boosted demand across the economy. As roads expanded and more people owned cars, cities spread outward, leading to rapid suburban growth. Improved transport also supported larger shopping centres, bigger stores and expanded entertainment options like movie theatres.
The trouble with the optimist’s argument
Now, the trouble is that even this is a slightly simplistic argument though it doesn’t sound like one. Let’s look at the Industrial Revolution, which started first in the United Kingdom and then moved to the United States. The first few decades in the late eighteenth and the early nineteenth century were tough for the workers.
As Abhijit Banerjee and Esther Duflo write in Good Economics for Hard Times: “Real blue-collar wages in Britain were almost halved between 1755 and 1802. Although 1802 was a particularly low year, they were on a declining trend between 1755 and the turn of the century, and it is only at the turn of the century that they started increasing again. They would recover their 1755 level only in 1820, sixty-five years later.”
This was because technology replaced workers. As Carl Benedikt Frey writes in The Technology Trap – Capital, Labour, and Power in the Age of Automation writes: “Artisan workers in the domestic system were replaced by machines, often tended by children – who had very little bargaining power and often worked without wages.”
This was an “era of intense deprivation and very difficult living conditions”.
Indeed, as Acemoglu and Johnson point out, by the mid-eighteenth century, the average person worked around 2,760 hours a year – a figure that had probably changed little over the previous century.
By 1800, annual working hours had climbed to 3,115. Over the next three decades, they increased further to 3,366 hours a year. Yet these longer hours did not translate into higher incomes for most people.
Of course, things eventually turned around, but by then three generations of Britons had seen their living standards decline. So, what happened?
In the first phase of the Industrial Revolution machines replaced humans, with new textile machinery replacing spinners and weavers. This put pressure on wages. The point being, when technology leads to automation, workers lose their existing jobs even as too few new productive roles are created to absorb them.
In the second phase of the Industrial Revolution, this began to change as technological progress shifted from replacing labour to enhancing it. As Acemoglu and Johnson write: “Arguably, the defining technology of the second half of the nineteenth century was railways.” Jobs were created as people worked on first building and then maintaining and running the railways.
A further wave of gains followed as abundant steel and cheaper coal fuelled the growth of other industries, including textiles and newer goods such as processed food, furniture and early household appliances. Railways also stimulated wholesale and retail trade.
The railways were followed by the spread of the telephone in the early twentieth century. Building and running the telephone systems proved to be labour intensive. It also got women into employment as telephone exchanges employed a large number of female workers.
Government involvement also played an important role. Across the Western world, the foundations of the welfare state were laid and education systems were strengthened. This mattered because as more factories were built, they needed skilled workers to serve as managers, accountants, clerks, salespeople, engineers and in other specialized roles.
This, as Frey writes, ensured that the “socialist revolution that Karl Marx prophesied did not happen” because “technology began to work in workers’ interests”. It can be said that governments and corporations read what Karl Marx prophesied and then worked towards making sure that that did not happen.
Further, the United States had abundant land and capital – that is a good amount of money – but it lacked skilled workers. As Acemoglu and Johnson write: “This high cost of skilled labour meant that American inventions often prioritized not just automation but also finding ways to boost the productivity of lower-skilled workers.” This helped in spreading prosperity, something that continued through much of the twentieth century across the Western world.
So, what does all this tell us?
First, techno optimism is a great selling proposition but like most simplistic arguments, it doesn’t really hold once you scratch the surface a little.
Second, new technologies lead to greater prosperity only when technological progress shifts from replacing labour to enhancing it. And that as we have seen can take time.
Third, from the time technology moves from replacing labour to working with it, a few decades can go by. Also, even if technology creates newer jobs faster, the ones losing their jobs to technology may not always be skilful enough to take on the newer jobs being created.
Now, learning new skills may take time, money and effort. Or as Susskind rhetorically argues: “Few lorry drivers in their fifties (if and when replaced by autonomous vehicles) will retrain as software engineers.” When we look at the issue from this lens, the entire argument that technology destroys jobs but creates new ones, isn’t really much of an argument.
Fourth, there is no way to know precisely what jobs the future will create. At the start of the Industrial Revolution, no one could have predicted that many would one day work as telegraph and telephone operators, locomotive engineers, or for that matter railway repairmen. So, despite all the techno optimism we really don’t know.
Fifth, as Joseph Schumpeter, the economist who came up with the term creative destruction to explain how free market capitalism works, wrote in Capitalism, Socialism and Democracy: “Any pro-capitalist argument must rest on long-run considerations…The unemployed of today would have completely to forget his personal fate.”
In today’s AI mahaul, such individual fates are dismissed as mere rounding errors.
Tech businessmen and other mahaul creators who they fund, promise a long-run productivity utopia to protect their massive bets, yet history – as we have seen above – warns that short-run overestimation often leaves real people stranded.
We risk overestimating the tool’s immediate magic while ignoring the human cost of the transition.
The model of shared prosperity has broken
Now, the Western world continued to see the benefits of the Industrial Revolution through a major part of the twentieth century as well. Indeed, as Acemoglu and Johnson point out: “Throughout most of the twentieth century, about 67-70 percent of national income went to workers, and the rest went to capital (in the form of payments for machinery and profits).”
But this started to change towards the end of 1970, when the workers’ share of the American national income was at 65%. In 2023, the latest data available it was at 57%. A rising share of national income now goes to corporate profits.
Indeed, what’s true about the US is also true about the other parts of the western world. So, the model of shared prosperity that led to the economic rise of the western world has largely broken down as economic power has become more concentrated in the hands of large corporations.
Why has this happened? As Frey writes: “Rising profits and the falling labour share is linked to the automation of routine middle-income jobs (such as those of machine operators, bookkeepers and mortgage underwriters) and the shift of labour into low-income service jobs (for example, those of janitors, waiters, and receptionists).”
Also, over the past 30 years, technological change has produced very few new jobs that do not demand a college degree. The rise of computer technologies has driven down the incomes of the middle class.
This shift is particularly visible in manufacturing. For example, many tasks in the body shop – such as painting, welding, precision work and several assembly jobs – have been automated using robots and specialised software.
As Wolf writes: “Manufacturing industry used to generate a very large number of relatively highly paid and secure jobs.” That isn’t happening anymore. At least, not in the rich western world.
Also, routine tasks have been automated. As David H Autor wrote in his 2013 research paper titled The Task Approach to Labour Markets : An Overview: “ The occupations that have contracted most rapidly as a share of total employment over the last three decades – in particular, clerical, administrative support, sales, production and operative positions – are reasonably well characterized as routine task-intensive: many of the core tasks of these occupations follow precise, carefully codified procedures.” This is only going to increase in the years to come.
Further, the impacts of GenAI aren’t expected to be gender neutral. A March 2026 research brief published by the International Labour Organisation points out: “Female-dominated occupations, such as business administration and clerical support, are almost twice as likely to be exposed to Gen AI as male-dominated ones such as construction, manufacturing and trade (29 versus 16 per cent). They also face much higher automation risk (16 percent for female versus 3 percent of male-dominated occupations).”
And if all that wasn’t enough, we live in a world where chief executives of companies are obsessed with driving up the value of their stock options. Given that, cutting labour costs is an important goal. Hence, there is a reluctance to share higher productivity gains with workers. And that also means, there has been a preference towards automation and a general reluctance to create new tasks for workers.
Indeed as Wolf writes: “The transformations of labour markets over the past four decades – deindustrialization, deunionization, declining participation, liberalization, and the rise of the ‘gig economy’”— is closely associated with the rise of ‘precarious' employment.”
Over and above this, as is well documented, in the last few decades the rich have grown richer. Indeed, the AI revolution is happening at such a point in time.
So,can we get to the point — will AI destroy jobs?
Yes. But as Susskind writes: “Most of the short-term claims currently being made by commentators about AI in the professions and white-collar work hugely overstate its likely impact.”
And that’s because, first, adoption takes time, and second, not all new technologies arrive at the same time. Given this, the impact will be incremental rather than big bang, something that lazy thinking and theorising often seems to suggest.
Indeed, this is likely to be true over the short run of the next few years, possibly till the end of this decade. This is entirely in line with Amara’s law, which states that we tend to overestimate the effect of a new technology in the short term. But the incremental impact will keep adding up as the years go by.
In the short-term, the automation of work that started with the rise of computers, is likely to continue. Something that is already being seen in India’s information technology companies, which are barely recruiting any new people. India’s top five IT companies added just 2,505 employees to their combined workforce in calendar year 2025, after having reduced their headcount by nearly the same number in 2024.
One possible explanation is that some of the routine coding and testing work traditionally done by junior engineers in IT companies is increasingly being assisted – or replaced – by AI tools.
Not just junior roles in information technology companies, but roles like data analysts, copywriters, translators, stockers, call centre agents, etc., are all under the risk of becoming irrelevant.
In fact, a study published by three Stanford University researchers suggests that early-career workers (ages 22-25) in the most AI-exposed occupations – such as customer service, accounting and software development – have experienced a 16 percent relative decline in employment.
Other than rendering youngsters unemployed this also leads to a paradox: With fewer entry-level roles, it becomes harder for the next generation to gain the human experience that once set people apart from machines.
Indeed, we may not have seen a sudden wave of mass firings across the economy because of AI – though there are exceptions – but something quieter is happening: companies are hiring fewer entry-level workers, slowly cutting off the first rung of the career ladder.
Of course, the Amara’s law also states that we tend to underestimate the effect of a technology in the long-term. What is likely to happen beyond the short-term? As Susskind writes: “In the medium term, we will see the boundaries between traditional professions break down as AI enables multidisciplinary service. In the long term, the main competition for professional firms will be AI empowered clients, organizations that can do much for themselves without referring work to external experts.”
Or as R Krishna Kumar, professor at the Indian Institute of Technology Madras, wrote in a recent column in The Hindu Business Line: “In sum, no level of hierarchy is safe. It is a fallacy to claim that only junior level programmers will be affected. Anyone who sits at a computer for a job is at risk of losing their seat.”
But won’t AI create newer job profiles as the second part of the Industrial Revolution did? Perhaps. Maybe, in a decade or two or more. But there are no guarantees here.
First, as Acemoglu and Johnson write: “Technology does not have a preordained direction, and nothing about it is inevitable.” As discussed earlier, the second part of the Industrial Revolution ended up creating jobs because we saw the advent of railways: a technology which was labour enhancing and not labour replacing. Also, governments got involved.
Second, as Banerjee and Duflo put it: “Nothing tells us the rebound is guaranteed to happen. There may well be no rebound from the fall in demand for labour resulting from this wave of automation and AI. Sectors that become more profitable may invest in new labour-saving technologies instead of hiring more workers.”
Third, the chief executives of today are expected to deliver shareholder value and not create jobs. So, all the techno optimism might work out well for the tech bros but not for the world at large.
As Susskind writes, “the undeniable current concentration of economic power enjoyed by a small number of technology companies (including Apple, Alphabet (Google), Amazon, Meta (Facebook), and Microsoft)” may ensure “that these and future AI corporations will be less concerned with ensuring prosperity for all than with returns for their shareholders”.
In short, AI risks making the rich richer and the poor poorer, widening the gap between them and pushing the world towards an even more two-tiered society.
Or as Eswar Prasad writes in The Doom Loop – Why the World Economic Order is Spiralling Into Disord writes: “Both AI and fintech could lead to greater concentration, especially if large, established firms co-opt them and use it to strengthen their market power.”
Of course, the tech-bros are not talking about all this and projecting an optimistic scenario, but then they are expected to do that.
Fourth, even if newer jobs might ultimately come by, it is possible that many people might have missed the bus by then, as was the case with the first phase of the Industrial Revolution. Further, as Susskind writes: “Our research suggested that most of these new jobs are likely themselves to be gobbled by technology.”
Fifth, in the very long-term, AI might arrive just in time to save the world. The population of large parts of the western world and countries like Japan and China are expected to shrink in the decades to come as women are having fewer babies than they did in the past.
In fact, in the Chinese case, the population is expected to shrink to 700 million by 2100 as against 1,400 million now. Given this, AI might just arrive at the right time replacing the decline in the labour force.
As Daniel Kahneman, the Nobel Prize-winning psychologist (he won the Economics prize), said in an interview: “You just don’t get it…In China, the robots are going to come just in time.”
Sixth, it has been argued that as long humans keep updating their skills they will continue to be in the game. Again, this is a very simplistic argument, for multiple reasons. Reskilling may not always be possible. It may not be quick enough. But most importantly, any reskilling can happen by taking into account what AI can do in a particular field at a given point of time. But then AI is constantly retraining itself. Given that, how does one take into account “the likely impact of not yet invented systems,” and train for that as well.
The new Ned Ludd
Given all this, it is worth asking whether workers around the world might eventually protest as AI begins to replace human labour – much as they did during the early years of the Industrial Revolution.
In 1811, a group of disgruntled textile workers in the United Kingdom, secretly organised and wrote to factory owners using the name “Ned Ludd,” a possibly mythical figure. In their letters, they warned that they would destroy mechanical knitting machines if their use continued. Their message resonated widely, and thousands of workers rallied behind the cause.
In 1811 and 1812, bands of textile workers began smashing machines that they believed threatened their livelihoods. These protesters came to be known as the Luddites.
As Banerjee and Duflo write: “Luddites destroyed machines to protest mechanization of weaving, which was threatening their livelihoods as skilled artisans. The term Luddite is now mostly used pejoratively to describe someone who blindly refuses progress.”
Now, will the rise of the AI lead to new Ned Ludds cropping up across different parts of the world?
Further, when the entry-level rungs of the career ladder are sawed off, do we really expect the youth to quietly “reskill” for jobs that haven’t been invented yet? Or will they simply lose faith in a system that views their first five years of professional growth as a “redundant cost”? Will they then become fodder for populist politicians as those in the western world who have lost out because of globalisation?
Will governments get involved in deciding the direction that AI takes? Or will they let companies keep taking AI where-ever they want to? At least that’s what the current evidence suggests, given that the biggest technology companies have been allowed to become monopolies in their respective areas.
Will the governments in the western world overlook the social and economic impacts of AI? Will governments allow “an unbalanced technology portfolio prioritizing automation and ignoring the creation of new tasks for workers” to continue? Will technology companies become feudalists?
Will we continue to let those who benefit most from the success of AI to set the agenda around it? Or as Susskind writes: “We need to rise above the homespun philosophizing that is popular in the tech community. Platitude is not principle.” But is that going to happen?
Indeed, as Adam Becker writes in More Everything Forever – AI Overlords, Space Empires, and Silicon Valley’s Crusade to Control the Fate of Humanity: “Setting the terms of such conversations about the future carries power in the present. If we don’t want tech billionaires setting those terms, we need to understand their ideas about the future.” Is that happening or will we continue to be mesmerised by tech billionaires and their agenda? Will they be the ones who decide if AI is working for the benefit of the people? Will they be the ones deciding on the morality of their creations?
If jobs are destroyed, and if enough new jobs are not created parallelly, what happens to consumption in the economy, especially in rich countries, which are largely consumption societies now? What happens to economic growth?
What happens to the psychological benefits of working? What about jobs giving a sense of purpose to the lived life?
What happens to taxes that governments collect? How does a state fund a welfare system when its primary tax-paying demographic is being replaced by lines of code?
Will inequality throughout the world increase as the two-tier society spreads further? If AI fails to lift all boats can we “take the broad acceptance of technological change for granted”?
Will, as Christopher Summerfield writes in These Strange New Minds “people whose jobs are automated… be left depending on the largesse of the state?”
These are important questions that aren’t really being asked, given that we seem to live in a world carried away by the hype around AI built up by tech bros.
So, where does all this leave us?
AI is often presented as an unstoppable force that will quickly transform the world of work. But the history of technology suggests a far more complicated reality. New technologies are usually surrounded by enormous hype in their early years. Nonetheless, adoption turns out to be slower and more uneven than the early predictions suggest.
This has been true of every technological revolution before this one – from the steam engine and railways to personal computers and the internet. Each of them eventually reshaped the economy, but the process took a while.
Indeed, AI is likely to follow a similar path. In the short run, its impact may be more gradual than the grand claims suggest. But over a longer period, its effects could turn out to be quite profound.
History offers a warning that we are not ready to hear. The early decades of the Industrial Revolution were marked by falling wages, longer working hours and severe hardship for most workers.
Prosperity arrived much later, and only after governments, institutions and businesses adjusted to the new economic reality – and only after a healthy fear of Karl Marx persuaded them that sharing the gains was preferable to revolution. Today, we have no equivalent fear.
The experience of the last few decades tells a similar story. Automation, computers and the internet dramatically improved productivity, but the gains were not evenly distributed. Routine middle-income jobs have been disappearing. High-skilled workers benefited disproportionately. Wages for many others stagnated. The result has been rising inequality and growing anxiety. AI could amplify all of these trends – and it is arriving at precisely the moment when the model of shared prosperity has already broken down.
Schumpeter was honest about what this means. Creative destruction, he wrote, requires the unemployed of today to “completely forget his personal fate”. In today’s mahaul, that forgetting is outsourced to a LinkedIn post about embracing change and a six-week online course in prompt engineering.
The questions that actually matter – what happens to consumption if the middle class continues to shrink, or what happens to tax revenues if white-collar work continues to be automated, or how a government can fund a welfare state if its primary taxpaying demographic gets replaced by lines of code – are nowhere on the agenda.
These questions are not being asked at the AI summits. They are not in the keynote speeches. They are not in the consultant reports. That is why the real debate about AI should not only be about technological possibilities. It should also be about who gains, who loses and how societies choose to manage the transition.
Some jobs will disappear, others will change, and new roles may eventually emerge. But there is no guarantee that the transition will be smooth or that the benefits will be widely shared.
As Morrissey sings: All make-up is a lie. When the make-up eventually comes off AI, we will discover that the future was not evenly distributed and that much of the hype may turn out to be just that. Nonetheless, as Emily M. Bender and Alex Hanna write in The AI Con – How to Fight Big Tech’s Hype and Create the Future We Want write: “Unfortunately, the perniciousness of hype is that it doesn’t need to be true to have huge impacts.”
At the end of the decades, we might live and we might just learn.
Vivek Kaul is an economic commentator and a writer.
This article has been republished from Newslaundry. You can read the original article here.