Technological unemployment facts for kids

In the 21st century, robots are starting to do jobs not just in factories but also in places like hospitals.

Technological unemployment happens when people lose their jobs because of new technology. It's a type of job loss where changes in technology mean fewer people are needed for certain tasks. This often involves machines that can do physical work ("mechanical-muscle") or smart computer programs that can do thinking tasks ("mechanical-mind"). These new technologies reduce the need for humans.

Think of how cars replaced horses for transport, or tractors replaced horses on farms. Similarly, human jobs have changed throughout history. For example, during the Industrial Revolution, many skilled weavers lost their jobs when new machines like the mechanized loom were invented. More recently, self-checkout machines and cashierless stores have replaced retail cashiers.

Most people agree that new technology can cause job losses for a short time. But whether it leads to long-term job losses has been a big debate. People in this debate are usually called "optimists" or "pessimists."

Optimists believe that even if technology causes short-term job losses, new jobs will always appear in the long run. Pessimists, however, think that sometimes new technologies can lead to a lasting drop in the total number of jobs. The term "technological unemployment" became popular in the 1930s by John Maynard Keynes, who thought it was just a "temporary problem." People have been talking about machines replacing human workers since at least Aristotle's time.

Before the 1700s, most people, rich or poor, worried about technology taking jobs. But this wasn't a big concern because there usually weren't many unemployed people. In the 1700s, as the Industrial Revolution grew, especially in Great Britain, fears about machines taking jobs increased. However, some thinkers started to say that new ideas wouldn't harm jobs overall. These ideas became clearer in the early 1800s with the classical economists. By the late 1800s, it seemed clear that technology was helping everyone, including working-class people. Worries about technology causing job losses faded. The idea that technology would cause lasting harm to jobs was even called the "Luddite fallacy."

Still, a few economists have always argued that technology could lead to long-term unemployment. For example, David Ricardo warned about this in the early 1800s. There were also many warnings in the 1930s and 1960s. In the late 1900s, especially in Europe, people worried again as unemployment stayed high in many industrial countries. But for most of the 1900s, most economists and the public believed that technology wouldn't cause long-term joblessness.

In the 2010s, new studies suggested that technological unemployment might increase worldwide. For example, Oxford professors Carl Benedikt Frey and Michael Osborne estimated that nearly half of U.S. jobs could be taken over by automation. However, their methods have been questioned. Even so, concern about technological unemployment is growing again. A 2017 Wired report quoted experts saying that dealing with job losses from automation is a "big issue." New technologies can now replace humans in many types of jobs, including professional, office, and creative roles. The World Bank's 2019 report says that while automation does replace workers, new technology also creates more new industries and jobs overall.

History of Job Changes

Ancient Times

Roman Emperor Vespasian did not allow a new, cheaper way to transport heavy goods. He wanted to protect jobs for workers.

The idea of technology causing job loss probably goes back to the invention of the wheel. Ancient societies had ways to help people who couldn't find work. In Ancient China and ancient Egypt, there might have been government programs to help with job loss from technology. In ancient Greece, many free workers could become unemployed because of new labor-saving tools or competition from slaves. Sometimes, these workers struggled greatly, but other times, they received help. Pericles started public works projects to give jobs to the unemployed.

One of the first thinkers to discuss this was Aristotle. He wrote in his book Politics that if machines became smart enough, humans wouldn't need to work anymore. In ancient Rome, people also helped the poor with handouts, like the Cura Annonae (grain dole). Sometimes, hundreds of thousands of families were supported this way. Less often, jobs were created through public works programs. Some emperors even refused or banned new inventions that saved labor. For example, Emperor Vespasian stopped a new way to transport heavy goods cheaply, saying, "You must allow my poor hauliers to earn their bread." By the late 200s AD, the Roman Empire started to have labor shortages, and widespread unemployment in Europe largely disappeared for over a thousand years.

Middle Ages and Renaissance

During the medieval and early Renaissance periods, many new technologies were adopted. Some were invented in Europe, while others came from places like China, India, and Arabia. The Black Death in the 1300s led to fewer workers across Europe. But by the 1400s, widespread unemployment started to return, especially in Western, Central, and Southern Europe. This was partly due to more people and changes in land use.

Because of the threat of job loss, people were less open to new technologies. European leaders often supported groups like Guilds, which banned new technologies. Sometimes, people who tried to promote or trade in these new inventions were even punished.

1500s to 1700s

Elizabeth I would not give a patent for a knitting machine. She said it would make her "poor subjects" lose their jobs and become beggars.

In Great Britain, leaders started to be less strict about new inventions earlier than in other parts of Europe. This might be why Britain led the Industrial Revolution. However, worries about technology and jobs remained strong in the 1500s and early 1600s.

A famous story is about William Lee, who invented a knitting machine that saved labor. He showed it to Queen Elizabeth I. But the Queen refused to give him a patent (a special right to his invention). She worried it would cause unemployment among textile workers. Lee tried to promote his invention in France and then returned to England, but Elizabeth's successor, James I, also refused for the same reason.

After the Glorious Revolution in the late 1600s, leaders became less sympathetic to workers' worries about losing jobs to new inventions. A new idea called Mercantilism became popular. It suggested that labor-saving technology would actually reduce unemployment by helping British companies sell more goods than foreign competitors. From the early 1700s, workers could no longer rely on government support against the threat of technological unemployment. They sometimes took direct action, like breaking machines, to protect their jobs.

Even though some thinkers like von Justi warned more often about technological unemployment, the main idea among leaders was that it wouldn't be a long-term problem.

1800s: Big Debates

The 1800s saw intense debates about technological unemployment, especially in Great Britain. New economic thinkers started to create the modern study of economics. Most of them agreed that technology wouldn't cause lasting unemployment.

However, in the early 1800s, some important economists, like Sismondi, Malthus, J. S. Mill, and even David Ricardo (from 1821), argued against this optimistic view. Ricardo was very respected, so his opinion was a challenge.

Jean-Baptiste Say argued that no one would use machines if they reduced production. He also believed that new supply creates its own demand, meaning displaced workers would find new jobs as the market adjusted. Ramsey McCulloch expanded on Say's ideas, supported by others like Charles Babbage.

Around the mid-1800s, Karl Marx joined the debates. He built on Ricardo's and Mill's ideas but was much more pessimistic about technological unemployment. His ideas gained many followers, but they didn't change mainstream economics much. By the 1870s, worries about technological unemployment faded in Britain. It became clear that new inventions were making everyone, including workers, more prosperous.

1900s: Ups and Downs

Some people argue that technology helps workers and doesn't replace them on a large scale.

In the early 1900s, widespread unemployment wasn't a big problem. While some thinkers continued to challenge the optimistic view, technological unemployment wasn't a major concern for most economists until the late 1920s. In the 1920s, high unemployment returned to Europe. Even in the U.S., urban unemployment started to rise in 1927. Farmers in America had been losing jobs since the early 1920s due to better farm technology like the tractor.

The main economic debates moved from Britain to the United States. Here, two big periods of debate about technological unemployment happened: in the 1930s and the 1960s. Both times, academic debates followed public concern about rising unemployment. Both debates ended when unemployment was reduced by wars: World War II for the 1930s debate, and the Vietnam War for the 1960s debate.

In the 1930s, optimists believed that markets would fix any short-term unemployment through "compensation effects" (new jobs appearing). In the 1960s, economists believed that government action could help if the market didn't fix the problem. Both times, major government studies found that long-term technological unemployment wasn't happening, though they agreed technology caused short-term job displacement and advised government help.

After the "golden age of capitalism" ended in the 1970s, unemployment rose again and stayed high in many rich countries for the rest of the century. Some economists, like Paul Samuelson, argued this might be due to new technology. Overall, in the late 1900s, most concern about technological unemployment was in Europe. Several popular books also warned about it, like The End of Work by Jeremy Rifkin. But for most of the 1900s, the general agreement among economists and the public was that technology does not cause long-term joblessness.

2000s: Renewed Concerns

The idea that new inventions don't cause long-term unemployment remained strong in the early 2000s. But some academic and popular books, like Robotic Nation by Marshall Brain and The Lights in the Tunnel by Martin Ford, continued to challenge it.

Since their 2011 book Race Against the Machine, MIT professors Andrew McAfee and Erik Brynjolfsson have been important voices raising concerns. They are still somewhat optimistic, saying that the key is not to compete against machines, but to compete with machines.

Concerns grew in 2013 because of studies predicting much more technological unemployment in the coming decades. There was also evidence that in some industries, jobs were falling worldwide even as production increased. This suggested that globalization wasn't the only cause of rising unemployment.

In 2014, the Financial Times reported that the impact of new inventions on jobs was a major topic in economic discussions. Former U.S. Treasury Secretary Lawrence Summers said in 2014 that he no longer believed automation would always create new jobs. He noted that more job sectors were losing jobs than creating new ones.

At the 2014 World Economic Forum meeting in Davos, many experts agreed that technology was leading to economic growth without creating many jobs. In 2015, Martin Ford won an award for his book Rise of the Robots: Technology and the Threat of a Jobless Future. Also in 2015, the first world summit on technological unemployment was held in New York. In 2016, US President Barack Obama said that due to artificial intelligence, society might debate "unconditional free money for everyone" within 10 to 20 years. In 2019, AI expert Stuart J. Russell said that "nearly all current jobs will go away" in the long run, so we need big policy changes.

However, other economists argue that long-term technological unemployment is unlikely. A 2014 survey of technology experts and economists found opinions split: 48% thought new technologies would displace more jobs than they created by 2025, while 52% disagreed. Professor Bruce Chapman advises that studies like Frey and Osborne's might overestimate job losses because they don't account for new jobs that technology will create in areas we don't even know about yet.

Studies on Job Risk

Many studies predict that automation will take a large number of jobs in the future, but the exact numbers vary. Research by Carl Benedikt Frey and Michael Osborne in 2013 showed that jobs with "well-defined procedures that can easily be done by smart computer programs" are at risk. Their study found that automation can affect both skilled and unskilled jobs, but low-paying physical jobs are most at risk. They estimated that 47% of U.S. jobs were at high risk.

Other studies using similar methods found:

In 2014, a study claimed 54% of jobs in the European Union were at risk.
A 2015 report found 41% of jobs in Israel were at risk.
A 2016 study found the risk was much higher in developing countries: 77% in China, 69% in India, 85% in Ethiopia, and 55% in Uzbekistan.
Another 2016 study found 74% of electrical & electronics jobs in Thailand and 81% in the Philippines were at high risk.
A 2016 United Nations report said 75% of jobs in the developing world were at risk.

The Council of Economic Advisers in the U.S. estimated in 2016 that 83% of jobs paying less than $20/hour were at risk, compared to only 4% of jobs paying over $40/hour. A 2017 study by PricewaterhouseCoopers found that up to 38% of jobs in the US, 35% in Germany, 30% in the UK, and 21% in Japan could be automated by the early 2030s.

However, not all recent studies agree that automation will cause widespread unemployment. A 2015 study on industrial robots in 17 countries found no overall job reduction and a slight increase in wages. A 2015 report by McKinsey Quarterly suggested that computers usually automate parts of tasks, rather than replacing entire jobs. A 2016 OECD study found that only about 9% of jobs in surveyed countries were in danger of automation, though this varied a lot by country. This study looked at more than just job tasks, including things like gender, education, and age.

A 2017 study on Germany found that automation didn't cause total job losses. Instead, jobs shifted from the industrial sector to the service sector. Manufacturing workers were not at risk and were more likely to stay employed, though their tasks might change. However, automation did lead to a decrease in workers' share of income, as it boosted productivity but not wages.

A 2018 study found that automation held down wages, even though it didn't reduce the total number of jobs. It found that automation reduced the share of value added by human labor, which slowed wage growth. In 2018, Adair Turner stated that 50% of jobs could already be automated with current technology, and all jobs could be automated by 2060.

Artificial Intelligence and Jobs

Since about 2017, there's been a new wave of concern about technological unemployment, this time focusing on artificial intelligence (AI). Some experts warn that if not managed, AI could cause an "economic singularity," where jobs change too fast for humans to keep up, leading to widespread unemployment. However, they also say that with the right actions from leaders and society, AI could be very good for workers.

Researchers note that it's hard to predict exactly what AI will do to jobs. Some argue that middle-class jobs, like professional services, are most likely to be fully replaced by AI. This could force workers to take lower-level jobs if they don't have the skills for high-level ones. However, others predict that AI will mostly "complement people" rather than "replicating people," meaning AI will help workers, not replace them. Studies also show that AI can create low-skill jobs, for example, training AI in low-income countries.

Many countries are now creating AI strategies, worried about falling behind in the "AI arms race." They hope that leading in AI will help their citizens get better jobs. For example, Finland offers a free online course on "The Elements of AI" in many European languages to help people gain skills. Oracle CEO Mark Hurd believes AI will create more jobs, not fewer, because humans will be needed to manage AI systems.

Martin Ford argues that many routine and predictable jobs could be automated in the next few decades. He warns that many new jobs might not be suitable for people with average abilities, even with retraining.

Some digital technologies, like modern robotics, have led to more jobs overall. But many businesses expect that automation will lead to job losses in the future, especially in Central and Eastern Europe. Other digital technologies, like platforms or big data, are expected to have a more neutral impact on jobs.

Key Ideas in the Debates

Long-Term Effects on Jobs

Everyone agrees that technology can cause temporary job losses and can also have good effects on workers. The main disagreement is whether technology can cause a lasting negative impact on the total number of jobs. Optimists believe that "compensation effects" will always create at least as many jobs as were lost. This optimistic view was dominant for most of the 1800s and 1900s. For example, some studies suggest that technology reduces unemployment in the long run.

The idea of structural unemployment (joblessness that doesn't go away even when the economy is strong) became popular in the 1960s. Pessimists see technological unemployment as a cause of this. Since the 1980s, even optimistic economists have agreed that structural unemployment has risen in rich countries. But they often blame it on globalisation and offshoring (moving jobs to other countries) rather than technology. In the 2000s, especially since 2013, pessimists are increasingly arguing that lasting worldwide technological unemployment is a growing threat.

Compensation Effects

John Kay inventor of the Fly Shuttle AD 1753, by Ford Madox Brown. This painting shows the inventor John Kay escaping an angry mob who were upset about his new machine. At this time, people didn't widely understand how new jobs could be created by technology.

Compensation effects are ways that new inventions can create jobs, making up for the jobs they initially destroy. In the 1820s, Jean-Baptiste Say described some of these effects. Later, Ramsey McCulloch developed a whole system of them. Karl Marx called this "compensation theory" and argued that these effects weren't guaranteed to work. Disagreement over how well compensation effects work has been central to the debates ever since.

Here are some compensation effects:

New machines: Jobs are created to build the new equipment needed for the innovation. (Economists rarely discuss this one now.)
New investments: Cost savings and increased profits from new technology can lead to new investments, creating jobs.
Changes in wages: If unemployment occurs, wages might lower, allowing more workers to be hired. Or, if workers' productivity rises, their wages might increase, leading to more spending and job creation.
Lower prices: New technology can lower prices, which leads to more demand for goods and services, and thus more jobs. Lower prices also help workers' buying power.
New products: Innovation can directly create entirely new jobs by leading to new products or services.

Even pessimists often agree that new products can have a positive effect on jobs. The debate continues about how well the other effects truly compensate workers for job losses. Some research suggests that new products create more jobs than new ways of making things (process innovation). Also, some studies show that for every skilled job created in high-tech industries, several other jobs are created in other sectors.

Many economists who are pessimistic about technological unemployment agree that compensation effects worked well in the 1800s and 1900s. But they argue that with the rise of computers, these effects are less effective. For example, Wassily Leontief argued in 1983 that early machines needed many human operators. But with computers, there's less need for human brainpower, which can lead to less pay and fewer jobs, even as productivity rises.

The Luddite Fallacy

The term "Luddite fallacy" is used to say that people who worry about long-term technological unemployment are wrong. It suggests they don't understand how compensation effects work. People who use this term usually believe that technology will not cause long-term unemployment and will eventually raise wages for everyone. The term comes from the Luddites, a group in the early 1800s who destroyed textile machines. For most of the 1900s and early 2000s, most economists believed that worrying about long-term technological unemployment was indeed a mistake. More recently, however, there's been more support for the idea that the benefits of automation are not shared equally.

There are two main ideas about why long-term job problems could happen:

The first idea, often linked to the Luddites, is that there's a limited amount of work available. If machines do it, there's nothing left for humans. Economists call this the "lump of labour fallacy" and argue that there's no such limit.
The second idea is that there's an infinite amount of work, but:

* Machines can do most of the "easy" work that doesn't need much skill or talent. * What counts as "easy" work keeps growing as technology gets better. * The work that's left might need more brainpower than most people have.

This second idea is supported by many who believe in the possibility of long-term technological unemployment today.

Skills and Technology

A common idea is that new technology mostly harms low-skilled workers but helps skilled workers. This might have been true for much of the 1900s. However, in the 1800s, new inventions often replaced expensive skilled craftspeople and generally helped low-skilled workers. In the 2000s, while some low-skilled jobs are being automated, other low-skilled jobs are still hard for machines to do. Meanwhile, office jobs requiring intermediate skills are increasingly being done by computer programs.

Some recent studies, however, found that industrial robots boost pay for highly skilled workers while having a negative impact on those with low to medium skills. Another report predicted that in the next ten years, automation would affect low-skilled workers the most.

Geoffrey Colvin at Forbes argued that predictions about what computers will never be able to do have often been wrong. He suggested that to know what skills humans will still need, we should look at tasks where we insist humans remain in charge, like judges or CEOs, or where human connection is essential, even if the tasks could be automated.

However, others believe even skilled human workers could become unnecessary. Oxford academics Frey and Osborne predicted that computerization could make nearly half of jobs redundant. They found a strong link between education and income and the ability to be automated, with office and service jobs being at higher risk. In 2012, Vinod Khosla, co-founder of Sun Microsystems, predicted that 80% of doctors' jobs would be lost to automated medical software in the next two decades.

A 2019 paper noted that in Russia, over 50% of workers do jobs that require low education and could be replaced by digital technologies. Only 13% of those people have an education level higher than current or near-future computer systems.

Solutions to Job Changes

Preventing Job Losses

Banning New Inventions

"What I object to, is the craze for machinery, not machinery as such. The craze is for what they call labour-saving machinery. Men go on 'saving labour', till thousands are without work and thrown on the open streets to die of starvation." — Gandhi, 1924

Historically, new inventions were sometimes banned because of worries about jobs. But with modern economics, this idea is generally not considered a solution for rich countries. Even those who worry about long-term technological unemployment usually see new inventions as good for society overall. J. S. Mill is one of the few Western economists who suggested banning technology as a solution.

Gandhian economics suggested delaying the use of labor-saving machines until unemployment was solved. However, this advice was mostly rejected by Nehru, who became India's prime minister. But slowing down new inventions to avoid technological unemployment was done in China under Mao in the 1900s.

Shorter Working Hours

In 1870, the average American worker worked about 75 hours a week. Before World War II, this fell to about 42 hours a week, and similar drops happened in other rich countries. Some economists, like Wassily Leontief, saw this as a choice to have more technological unemployment. Reducing working hours helped share the available work. Workers were happy to work less for more leisure, as technology was generally increasing their pay.

Further reductions in working hours have been suggested as a solution by economists like John R. Commons and John Maynard Keynes. However, once working hours reached about 40 hours per week, workers were less keen on further reductions. This was partly to avoid losing income and partly because many value work itself. In the 1900s, economists generally argued against further reductions as a solution to unemployment, saying it was based on the "lump of labor fallacy." In 2014, Google's co-founder, Larry Page, suggested a four-day workweek to help more people find jobs as technology replaces them.

Public Works

Public works programs (government projects like building roads or bridges) have traditionally been used to create jobs directly. Some people, like professor Mathew Forstater, believe that public works and guaranteed government jobs might be the best solution to technological unemployment. Unlike welfare, these programs give people the social recognition and meaning that comes with work.

For developing countries, public works might be easier to manage than universal welfare programs. Even economists who usually support free markets, like Larry Summers, have recommended spending on infrastructure (like energy and transportation) as a solution to technological unemployment.

Education

Improving access to good education, including skills training for adults, is a solution that almost everyone agrees on. It's especially popular with businesses. However, some experts argue that education alone won't be enough to solve technological unemployment. They point out that demand for many intermediate skills has dropped, and not everyone can become an expert in the most advanced skills. Paul Krugman argued in 2011 that better education would not be enough.

Living with Technological Unemployment

Welfare Payments

Using different types of government aid has often been accepted as a solution to technological unemployment, even by those who are optimistic about jobs in the long run. Welfare programs have tended to last longer than other solutions like direct job creation. Even though he developed the idea of compensation effects, Ramsey McCulloch and most other classical economists supported government aid for those affected by technological unemployment. They understood that the market doesn't adjust instantly, and displaced workers can't always find new jobs right away.

Basic Income

Many people have argued that traditional welfare payments might not be enough for the future challenges of technological unemployment. They suggest a basic income as an alternative. A basic income is a regular payment from the government to everyone, regardless of whether they work. People like Martin Ford, Erik Brynjolfsson, Robert Reich, Andrew Yang, and Elon Musk support some form of basic income. Reich has said it's "almost inevitable."

Since late 2015, new basic income trials have started in Finland, the Netherlands, and Canada. Many technology entrepreneurs also support basic income.

Some people are skeptical about basic income. One concern is that it might make people not want to work. However, evidence from older trials in India, Africa, and Canada suggests that this doesn't happen. Instead, a basic income can encourage small businesses and more productive, cooperative work. Another big challenge is how to pay for it. New ideas for funding have been proposed, but how to pay for a generous basic income remains a debated question.

Broadening Ownership of Technology

Some solutions don't fit easily into traditional political ideas. One is to spread the ownership of robots and other productive assets (things that produce wealth) more widely. People like James S. Albus and Richard B. Freeman have supported this idea.

Jaron Lanier suggested a similar solution: a system where ordinary people receive "nano payments" for the big data they create by using the internet.

Changes Towards a Post-Scarcity Economy

Some groups propose big structural changes towards a "post-scarcity economy." In this future, people are "freed" from boring, automatable jobs instead of "losing" them. In such a system, all jobs would either be automated, removed if they don't create real value (like some advertising), or done based on helping others and society, rather than for money. These groups also believe that the free time people gain would lead to a boom in creativity, invention, and community, while reducing stress.

Other Ideas

Sometimes, the threat of technological unemployment has been used by economists who support free markets to argue for changes that make it easier for employers to hire and fire workers. On the other hand, it has also been used to argue for more protection for employees.

Economists like Larry Summers have suggested that a mix of solutions might be needed. He advised strong efforts to stop wealthy people from avoiding taxes and to make it harder to get rich without making "great social contributions." Summers also suggested stronger anti-monopoly laws, less protection for intellectual property, more encouragement for profit-sharing with workers, better company management, stronger financial rules, easier land-use rules, better training for young people and displaced workers, and more public and private investment in infrastructure.

Michael Spence has advised that understanding global forces and technology will be key to responding to its future impact. Adapting will require changes in thinking, policies, investments (especially in human capital or people's skills), and possibly new ways of working and sharing wealth.