Special report: - Rollins College  · Web viewEven after computers beat grandmasters at chess ... 500 company and one of the world’s biggest white-goods ... radical transparency

The Economist

The future of jobs

The onrushing wave

Previous technological innovation has always delivered more long-run employment, not less. But things can changeJan 18th 2014

IN 1930, when the world was “suffering…from a bad attack of economic pessimism”, John Maynard Keynes wrote a broadly optimistic essay, “Economic Possibilities for our Grandchildren”. It imagined a middle way between revolution and stagnation that would leave the said grandchildren a great deal richer than their grandparents. But the path was not without dangers.

One of the worries Keynes admitted was a “new disease”: “technological unemployment…due to our discovery of means of economising the use of labour outrunning the pace at which we can find new uses for labour.” His readers might not have heard of the problem, he suggested—but they were certain to hear a lot more about it in the years to come.

For the most part, they did not. Nowadays, the majority of economists confidently wave such worries away. By raising productivity, they argue, any automation which economises on the use of labour will increase incomes. That will generate demand for new products and services, which will in turn create new jobs for

displaced workers. To think otherwise has meant being tarred a Luddite—the name taken by 19th-century textile workers who smashed the machines taking their jobs.

For much of the 20th century, those arguing that technology brought ever more jobs and prosperity looked to have the better of the debate. Real incomes in Britain scarcely doubled between the beginning of the common era and 1570. They then tripled from 1570 to 1875. And they more than tripled from 1875 to 1975. Industrialisation did not end up eliminating the need for human workers. On the contrary, it created employment opportunities sufficient to soak up the 20th century’s exploding population. Keynes’s vision of everyone in the 2030s being a lot richer is largely achieved. His belief they would work just 15 hours or so a week has not come to pass.

When the sleeper wakes

Yet some now fear that a new era of automation enabled by ever more powerful and capable computers could work out differently. They start from the observation that, across the rich world, all is far from well in the world of work. The essence of what they see as a work crisis is that in rich countries the wages of the typical worker, adjusted for cost of living, are stagnant. In America the real wage has hardly budged over the past four decades. Even in places like Britain and Germany, where employment is touching new highs, wages have been flat for a decade. Recent research suggests that this is because substituting capital for labour through automation is increasingly attractive; as a result owners of capital have captured ever more of the world’s income since the 1980s, while the share going to labour has fallen.

At the same time, even in relatively egalitarian places like Sweden, inequality among the employed has risen sharply, with the share going to the highest earners soaring. For those not in the elite, argues David Graeber, an anthropologist at the London School of Economics, much of modern labour consists of stultifying “bullshit jobs”—low- and mid-level screen-sitting that serves simply to occupy workers for whom the economy no longer has much use. Keeping them employed, Mr Graeber argues, is not an economic choice; it is something the ruling class does to keep control over the lives of others.

Be that as it may, drudgery may soon enough give way to frank unemployment. There is already a long-term trend towards lower levels of employment in some rich countries. The proportion of American adults participating in the labour force recently hit its lowest level since 1978, and although some of that is due to the effects of ageing, some is not. In a recent speech that was modelled in part on Keynes’s “Possibilities”, Larry Summers, a former American treasury secretary, looked at employment trends among American men between 25 and 54. In the 1960s only one in 20 of those men was not working. According to Mr Summers’s extrapolations, in ten years the number could be one in seven.

This is one indication, Mr Summers says, that technical change is increasingly taking the form of “capital that effectively substitutes for labour”. There may be a lot more for such capital to do in the near future. A 2013 paper by Carl Benedikt Frey and Michael Osborne, of the University of Oxford, argued that jobs are at high risk of being automated in 47% of the occupational categories into which work is customarily sorted. That includes accountancy, legal work, technical writing and a lot of other white-collar occupations.

Answering the question of whether such automation could lead to prolonged pain for workers means taking a close look at past experience, theory and technological trends. The picture suggested by this evidence is a complex one. It is also more worrying than many economists and politicians have been prepared to admit.

The lathe of heaven

Economists take the relationship between innovation and higher living standards for granted in part because they believe history justifies such a view. Industrialisation clearly led to enormous rises in incomes and living standards over the long run. Yet the road to riches was rockier than is often appreciated.

In 1500 an estimated 75% of the British labour force toiled in agriculture. By 1800 that figure had fallen to 35%. When the shift to manufacturing got under way during the 18th century it was overwhelmingly done at small scale, either within the home or in a small workshop; employment in a large factory was a rarity. By the end of the 19th century huge plants in massive industrial cities were the norm. The great shift was made possible by automation and steam engines.

Industrial firms combined human labour with big, expensive capital equipment. To maximise the output of that costly machinery, factory owners reorganised the processes of production. Workers were given one or a few repetitive tasks, often making components of finished products rather than whole pieces. Bosses imposed a tight schedule and strict worker discipline to keep up the productive pace. The Industrial Revolution was not simply a matter of replacing muscle with steam; it was a matter of reshaping jobs themselves into the sort of precisely defined components that steam-driven machinery needed—cogs in a factory system.

The way old jobs were done changed; new jobs were created. Joel Mokyr, an economic historian at Northwestern University in Illinois, argues that the more intricate machines, techniques and supply chains of the period all required careful tending. The workers who provided that care were well rewarded. As research by Lawrence Katz, of Harvard University, and Robert Margo, of Boston University, shows, employment in manufacturing “hollowed out”. As employment grew for highly skilled workers and unskilled workers, craft workers lost out. This was the loss to which the Luddites, understandably if not effectively, took exception.

With the low-skilled workers far more numerous, at least to begin with, the lot of the average worker during the early part of this great industrial and social upheaval was not a happy one. As Mr Mokyr notes, “life did not improve all that much between 1750 and 1850.” For 60 years, from 1770 to 1830, growth in British wages, adjusted for inflation, was imperceptible because productivity growth was restricted to a few industries. Not until the late 19th century, when the gains had spread across the whole economy, did wages at last perform in line with productivity (see chart 1).

Along with social reforms and new political movements that gave voice to the workers, this faster wage growth helped spread the benefits of industrialisation across wider segments of the population. New investments in education provided a supply of workers for the more skilled jobs that were by then being created in ever greater numbers. This shift continued into the 20th century as post-secondary education became increasingly common.

Claudia Goldin, an economist at Harvard University, and Mr Katz have written that workers were in a “race between education and technology” during this period, and for the most part they won. Even so, it was not until the “golden age” after the second world war that workers in the rich world secured real prosperity, and a large, property-owning middle class came to dominate politics. At the same time communism, a legacy of industrialisation’s harsh early era, kept hundreds of millions of people around the world in poverty, and the effects of the imperialism driven by European industrialisation continued to be felt by billions.

The impacts of technological change take their time appearing. They also vary hugely from industry to industry. Although in many simple economic models technology pairs neatly with capital and labour to produce output, in practice technological changes do not affect all workers the same way. Some find that their skills are complementary to new technologies. Others find themselves out of work.

Take computers. In the early 20th century a “computer” was a worker, or a room of workers, doing mathematical calculations by hand, often with the end point of one person’s work the starting point for the next. The development of mechanical and electronic computing rendered these arrangements obsolete. But in time it greatly increased the productivity of those who used the new computers in their work.

Many other technical innovations had similar effects. New machinery displaced handicraft producers across numerous industries, from textiles to metalworking. At the same time it enabled vastly more output per person than craft producers could ever manage.

Player piano

For a task to be replaced by a machine, it helps a great deal if, like the work of human computers, it is already highly routine. Hence the demise of production-line jobs and some sorts of book-keeping, lost to the robot and the spreadsheet. Meanwhile work less easily broken down into a series of stereotyped tasks—whether rewarding, as the management of other workers and the teaching of toddlers can be, or more of a grind, like tidying and cleaning messy work places—has grown as a share of total employment.

But the “race” aspect of technological change means that such workers cannot rest on their pay packets. Firms are constantly experimenting with new technologies and production processes. Experimentation with different techniques and business models requires flexibility, which is one critical advantage of a human worker. Yet over time, as best practices are worked out and then codified, it becomes easier to break production down into routine components, then automate those components as technology allows.

If, that is, automation makes sense. As David Autor, an economist at the Massachusetts Institute of Technology (MIT), points out in a 2013 paper, the mere fact that a job can be automated does not mean that it will be; relative costs also matter. When Nissan produces cars in Japan, he notes, it relies heavily on robots. At plants in India, by contrast, the firm relies more heavily on cheap local labour.

Even when machine capabilities are rapidly improving, it can make sense instead to seek out ever cheaper supplies of increasingly skilled labour. Thus since the 1980s (a time when, in America, the trend towards post-secondary education levelled off) workers there and elsewhere have found themselves facing increased competition from both machines and cheap emerging-market workers.

Such processes have steadily and relentlessly squeezed labour out of the manufacturing sector in most rich economies. The share of American employment in manufacturing has declined sharply since the 1950s, from almost 30% to less than 10%. At the same time, jobs in services soared, from less than 50% of employment to almost 70% (see chart 2). It was inevitable, therefore, that firms would start to apply the same experimentation and reorganisation to service industries.

A new wave of technological progress may dramatically accelerate this automation of brain-work. Evidence is mounting that rapid technological progress, which accounted for the long era of rapid productivity growth from the 19th century to the 1970s, is back. The sort of advances that allow people to put in their pocket a computer that is not only more powerful than any in the world 20 years ago, but also has far better software and far greater access to useful data, as well as to other people and machines, have implications for all sorts of work.

The case for a highly disruptive period of economic growth is made by Erik Brynjolfsson and Andrew McAfee, professors at MIT, in “The Second Machine Age”, a book to be published later this month. Like the first great era of industrialisation, they argue, it should deliver enormous benefits—but not without a period of disorienting and uncomfortable change. Their argument rests on an underappreciated aspect of the exponential growth in chip processing speed, memory capacity and other computer metrics: that the amount of progress computers will make in the next few years is always equal to the progress they have made since the very beginning. Mr Brynjolfsson and Mr McAfee reckon that the main bottleneck on innovation is the time it takes society to sort through the many combinations and permutations of new technologies and business models.

A startling progression of inventions seems to bear their thesis out. Ten years ago technologically minded economists pointed to driving cars in traffic as the sort of human accomplishment that computers were highly unlikely to master. Now Google cars are rolling round California driver-free no one doubts such mastery is possible, though the speed at which fully self-driving cars will come to market remains hard to guess.

Brave new world

Even after computers beat grandmasters at chess (once thought highly unlikely), nobody thought they could take on people at free-form games played in natural language. Then Watson, a pattern-recognising supercomputer developed by IBM, bested the best human competitors in America’s popular and syntactically tricksy general-knowledge quiz show “Jeopardy!” Versions of Watson are being marketed to firms across a range of industries to help with all sorts of pattern-recognition problems. Its acumen will grow, and its costs fall, as firms learn to harness its abilities.

The machines are not just cleverer, they also have access to far more data. The combination of big data and smart machines will take over some occupations wholesale; in others it will allow firms to do more with fewer workers. Text-mining programs will displace professional jobs in legal services. Biopsies will be analysed more efficiently by image-processing software than lab technicians. Accountants may follow travel agents and tellers into the unemployment line as tax software improves. Machines are already turning basic sports results and financial data into good-enough news stories.

Jobs that are not easily automated may still be transformed. New data-processing technology could break “cognitive” jobs down into smaller and smaller tasks. As well as opening the way to eventual automation this could reduce the satisfaction from such work, just as the satisfaction of making things was reduced by deskilling and interchangeable parts in the 19th century. If such jobs persist, they may engage Mr Graeber’s “bullshit” detector.

Being newly able to do brain work will not stop computers from doing ever more formerly manual labour; it will make them better at it. The designers of the latest generation of industrial robots talk about their creations as helping workers rather than replacing them; but there is little doubt that the technology will be able to do a bit of both—probably more than a bit. A taxi driver will be a rarity in many places by the 2030s or 2040s. That sounds like bad news for journalists who rely on that most reliable source of local knowledge and prejudice—but will there be many journalists left to care? Will there be airline pilots? Or traffic cops? Or soldiers?

There will still be jobs. Even Mr Frey and Mr Osborne, whose research speaks of 47% of job categories being open to automation within two decades, accept that some jobs—especially those currently associated with high levels of education and high wages—will survive (see table). Tyler Cowen, an economist at George Mason University and a much-read blogger, writes in his most recent book, “Average is Over”, that rich economies seem to be bifurcating into a small group of workers with skills highly complementary with machine intelligence, for whom he has high hopes, and the rest, for whom not so much.

And although Mr Brynjolfsson and Mr McAfee rightly point out that developing the business models which make the best use of new technologies will involve trial and error and human flexibility, it is also the case that the second machine age will make such trial and error easier. It will be shockingly easy to launch a startup, bring a new product to market and sell to billions of global consumers (see article). Those who create or invest in blockbuster ideas may earn unprecedented returns as a result.

In a forthcoming book Thomas Piketty, an economist at the Paris School of Economics, argues along similar lines that America may be pioneering a hyper-unequal economic model in which a top 1% of capital-owners and “supermanagers” grab a growing share of national income and accumulate an increasing concentration of national wealth. The rise of the middle-class—a 20th-century innovation—was a hugely important political and social development across the world. The squeezing out of that class could generate a more antagonistic, unstable and potentially dangerous politics.

The potential for dramatic change is clear. A future of widespread technological unemployment is harder for many to accept. Every great period of innovation has produced its share of labour-market doomsayers, but technological progress has never previously failed to generate new employment opportunities.

The productivity gains from future automation will be real, even if they mostly accrue to the owners of the machines. Some will be spent on goods and services—golf instructors, household help and so on—and most of the rest invested in firms that are seeking to expand and presumably hire more labour. Though inequality could soar in such a world, unemployment would not necessarily spike. The current doldrum in wages may, like that of the early industrial era, be a temporary matter, with the good times about to roll (see chart 3).

These jobs may look distinctly different from those they replace. Just as past mechanisation freed, or forced, workers into jobs requiring more cognitive dexterity, leaps in machine intelligence could create space for people to specialise in more emotive occupations, as yet unsuited to machines: a world of artists and therapists, love counsellors and yoga instructors.

Such emotional and relational work could be as critical to the future as metal-bashing was in the past, even if it gets little respect at first. Cultural norms change slowly. Manufacturing jobs are still often treated as “better”—in some vague, non-pecuniary way—than paper-pushing is. To some 18th-century observers, working in the fields was inherently more noble than making gewgaws.

But though growth in areas of the economy that are not easily automated provides jobs, it does not necessarily help real wages. Mr Summers points out that prices of things-made-of-widgets have fallen remarkably in past decades; America’s Bureau of Labour Statistics reckons that today you could get the equivalent of an early 1980s television for a twentieth of its then price, were it not that no televisions that poor are still made. However, prices of things not made of widgets, most notably college education and health care, have shot up. If people lived on widgets alone— goods whose costs have fallen because of both globalisation and technology—there would have been no pause in the increase of real wages. It is the

increase in the prices of stuff that isn’t mechanised (whose supply is often under the control of the state and perhaps subject to fundamental scarcity) that means a pay packet goes no further than it used to.

So technological progress squeezes some incomes in the short term before making everyone richer in the long term, and can drive up the costs of some things even more than it eventually increases earnings. As innovation continues, automation may bring down costs in some of those stubborn areas as well, though those dominated by scarcity—such as houses in desirable places—are likely to resist the trend, as may those where the state keeps market forces at bay. But if innovation does make health care or higher education cheaper, it will probably be at the cost of more jobs, and give rise to yet more concentration of income.

The machine stops

Even if the long-term outlook is rosy, with the potential for greater wealth and lots of new jobs, it does not mean that policymakers should simply sit on their hands in the mean time. Adaptation to past waves of progress rested on political and policy responses. The most obvious are the massive improvements in educational attainment brought on first by the institution of universal secondary education and then by the rise of university attendance. Policies aimed at similar gains would now seem to be in order. But as Mr Cowen has pointed out, the gains of the 19th and 20th centuries will be hard to duplicate.

Boosting the skills and earning power of the children of 19th-century farmers and labourers took little more than offering schools where they could learn to read, write and do algebra. Pushing a large proportion of college graduates to complete graduate work successfully will be harder and more expensive. Perhaps cheap and innovative online education will indeed make new attainment possible. But as Mr Cowen notes, such programmes may tend to deliver big gains only for the most conscientious students.

Another way in which previous adaptation is not necessarily a good guide to future employment is the existence of welfare. The alternative to joining the 19th-century industrial proletariat was malnourished deprivation. Today, because of measures introduced in response to, and to some extent on the proceeds of, industrialisation, people in the developed world are provided with unemployment benefits, disability allowances and other forms of welfare. They are also much more likely than a bygone peasant to have savings. This means that the “reservation wage”—the wage below which a worker will not accept a job—is now high in historical terms. If governments refuse to allow jobless workers to fall too far below the average standard of living, then this reservation wage will rise steadily, and ever more workers may find work unattractive. And the higher it rises, the greater the incentive to invest in capital that replaces labour.

Everyone should be able to benefit from productivity gains—in that, Keynes was united with his successors. His worry about technological unemployment was mainly a worry about a “temporary phase of maladjustment” as society and the economy adjusted to ever greater levels of productivity. So it could well prove. However, society may find itself sorely tested if, as seems possible, growth and innovation deliver handsome gains to the skilled, while the rest cling to dwindling employment opportunities at stagnant wages.

Robots in the rustbelt

The future lies in automation

Factories are upgrading, but still lag far behind the rich worldApr 8th 2017

WONG CHAP WING, a native of Hong Kong, runs a factory in Dongguan, an industrial city north of Shenzhen. Hip Fai, his privately held firm, stamps metal parts for things like printers and copiers. The energetic septuagenarian started dye- and mould-making in 1966, and recalls a time when migrants were grateful for a job. “There are not enough technical workers now,” he complains. Young people turn up their noses at factory work. He used to pay 600 yuan a month, but now they demand 5,000.

The future is not bright for workshops that cannot upgrade. Mr Wong looked into shifting to a cheaper location inland but decided that the savings were too small. He says that many low-end subcontractors in his area are closing down. Looking at the antiquated equipment and the throngs of workers in his factory, it seems this greasy and noisy place, too, may face extinction.

Turn a corner, though, and you spot the future: a hybrid assembly line where shiny Japanese robots are mingling with human workers. Peter Guarraia of Bain, a consultancy, explains that the big global trend in factory automation is “co-bots”: robots designed to collaborate safely with workers. They will look out for people and can be programmed by line workers.

Dongguan has an official policy of encouraging automation, part of a national strategy to upgrade manufacturing

Mr Wong spent 200,000 yuan on each robot but expects to get his money back within three years because his reconfigured assembly line is much more productive. Looking back, “I could not imagine my factory full of robots,” he reflects. “I came here for the cheap labour.”

Dongguan has an official policy of encouraging automation, and has set aside 200m yuan a year to help its factories eliminate jobs. This is part of a national strategy to upgrade manufacturing through automation. The governments of the PRD are leading the charge. Guangdong has pledged to spend 943bn yuan to boost

the manufacture and adoption of robotics in the province. Guangzhou optimistically hopes to automate the jobs of four-fifths of the city’s industrial workforce by 2020.

The productivity imperative

The sprawling headquarters of Midea in Foshan, a city near Guangzhou, look as though that day has already come. The firm was started in 1968 with 5,000 yuan, operating from a workshop measuring just 20 square metres. He Xiangjian, the founder, and his team scrounged what they could from Mao’s tattered economy to make plastic bottle caps, glass bottles and rubber balls. Today Midea is a Fortune 500 company and one of the world’s biggest white-goods manufacturers, selling everything from internet-controlled kitchen appliances to smart washing machines. Mr He, who retains a controlling stake in the firm, is a multi-billionaire. Last year Midea gobbled up Kuka, a German robotics firm, in a deal worth nearly $5bn. It also has a joint venture with Yaskawa, a Japanese robotics outfit. It is spending 10bn yuan to develop robots, both to use in its own factories and to sell to others.

There are two main reasons to think the delta’s factories need to upgrade. First, the level of automation in China remains low compared with some of its competitors. In 2015 the average for the country as a whole was fewer than 50 robots per 10,000 factory workers, compared with about 300 in Germany and Japan and more than 500 in South Korea (see chart).

Second, China’s supply of cheap labour is running out, which is pushing up wages steeply. China’s low birth rate, exacerbated by its one-child policy (now revoked), has meant that the working-age population has already peaked and is set to shrink significantly in the next few decades. The mass migration of poor rural dwellers from interior provinces to the PRD is slowing, and without that influx of labour, growth targets will be harder to hit.

As a consequence, China urgently needs to beef up its productivity. Over the two decades to 2016, labour productivity has risen by an average of 8.5% a year, but in the past three years this growth has slowed to less than 7% a year, and the absolute level remains low, at only 15-30% of that in OECD countries.

Yet automation should be market-driven, not subsidy-induced, and there are signs of a bubble. Thanks to the official push for “indigenous innovation”, Chinese automation firms are often subsidised even if their technology is not up to scratch.

In an era of rapid growth and cheap labour, Chinese bosses set up factories without much concern for efficiency or quality of tooling. If a problem arose, they would throw more men at the job rather than invest even in simple automation. Now many of them are uncritically replacing humans with hardware. AlixPartners, a consultancy, warns that China risks being “left behind as a failed low-cost-country-model economy”.

Karel Eloot of McKinsey, a consultancy, reckons that most Chinese firms are not even bothering to adopt such global best practices as Six Sigma, which uses statistical methods to ensure quality, and lean manufacturing, which emphasises efficiency and waste reduction. By one estimate, such tools could boost productivity by 15-30%. Instead, many firms are deploying robots to automate their current inefficient ways of working. Mr Eloot would like to see more data, measurement and analysis on the shop floor, with the lessons integrated into work routines.

That may sound too sophisticated, but the PRD’s firms are already showing the rest of China how to leapfrog on smart automation. Consider Ash Cloud’s factory in Shenzhen. This private company makes cheap plastic cases for mobile phones, each costing a few yuan. It sells about 35m of them a year, earning it about $35m in revenues. Although this is a brutally competitive niche, the firm’s profit margin is 10%.

Fred Chen, its general manager, reveals his secret: “Most Chinese firms suffer from production losses, mistakes, scrap, communications and production errors, warehouse mismanagement and so on…our success is due to very good controls.” The firm’s genius is in its manufacturing management system. Every employee has access to it from scores of iPads found all over the factory. There are cameras and sensors everywhere. The iPads display in large type how much net revenue has been earned from each product during a given shift.

A manager explains the advantages: “We have no information islands…radical transparency means no secrets, no turf battles.” Since everybody sees the data in real time, all can change plans on the fly. For Mr Chen the conclusion is obvious: “It is time for Chinese factories to change their management habits.”

Special report: Artificial intelligence

The impact on jobs

Automation and anxiety

Will smarter machines cause mass unemployment?Jun 25th 2016

SITTING IN AN office in San Francisco, Igor Barani calls up some medical scans on his screen. He is the chief executive of Enlitic, one of a host of startups applying deep learning to medicine, starting with the analysis of images such as X-rays and CT scans. It is an obvious use of the technology. Deep learning is renowned for its superhuman prowess at certain forms of image recognition; there are large sets of labelled training data to crunch; and there is tremendous potential to make health care more accurate and efficient.

Dr Barani (who used to be an oncologist) points to some CT scans of a patient’s lungs, taken from three different angles. Red blobs flicker on the screen as Enlitic’s deep-learning system examines and compares them to see if they are blood vessels, harmless imaging artefacts or malignant lung nodules. The system ends up highlighting a particular feature for further investigation. In a test against three expert human radiologists working together, Enlitic’s system was 50% better at classifying malignant tumours and had a false-negative rate (where a cancer is missed) of zero, compared with 7% for the humans. Another of Enlitic’s systems, which examines X-rays to detect wrist fractures, also handily outperformed human experts. The firm’s technology is currently being tested in 40 clinics across Australia.

A computer that dispenses expert radiology advice is just one example of how jobs currently done by highly trained white-collar workers can be automated, thanks to the advance of deep learning and other forms of artificial intelligence. The idea that manual work can be carried out by machines is already familiar; now ever-smarter machines can perform tasks done by information workers, too. What determines vulnerability to automation, experts say, is not so much whether the work concerned is manual or white-

collar but whether or not it is routine. Machines can already do many forms of routine manual labour, and are now able to perform some routine cognitive tasks too. As a result, says Andrew Ng, a highly trained and specialised radiologist may now be in greater danger of being replaced by a machine than his own executive assistant: “She does so many different things that I don’t see a machine being able to automate everything she does any time soon.”

So which jobs are most vulnerable? In a widely noted study published in 2013, Carl Benedikt Frey and Michael Osborne examined the probability of computerisation for 702 occupations and found that 47% of workers in America had jobs at high risk of potential automation. In particular, they warned that most workers in transport and logistics (such as taxi and delivery drivers) and office support (such as receptionists and security guards) “are likely to be substituted by computer capital”, and that many workers in sales and services (such as cashiers, counter and rental clerks, telemarketers and accountants) also faced a high risk of computerisation. They concluded that “recent developments in machine learning will put a substantial share of employment, across a wide range of occupations, at risk in the near future.” Subsequent studies put the equivalent figure at 35% of the workforce for Britain (where more people work in creative fields less susceptible to automation) and 49% for Japan.

What determines vulnerability to automation is not so much whether the work concerned is manual or white-collar but whether or not it is routine

Economists are already worrying about “job polarisation”, where middle-skill jobs (such as those in manufacturing) are declining but both low-skill and high-skill jobs are expanding. In effect, the workforce bifurcates into two groups doing non-routine work: highly paid, skilled workers (such as architects and senior managers) on the one hand and low-paid, unskilled workers (such as cleaners and burger-flippers) on the other. The stagnation of median wages in many Western countries is cited as evidence that automation is already having an effect—though it is hard to disentangle the impact of offshoring, which has also moved many routine jobs (including manufacturing and call-centre work) to low-wage countries in the developing world. Figures published by the Federal Reserve Bank of St Louis show that in America, employment in non-routine cognitive and non-routine manual jobs has grown steadily since the 1980s, whereas employment in routine jobs has been broadly flat (see chart). As more jobs are automated, this trend seems likely to continue.

And this is only the start. “We are just seeing the tip of the iceberg. No office job is safe,” says Sebastian Thrun, an AI professor at Stanford known for his work on self-driving cars. Automation is now “blind to the colour of your collar”, declares Jerry Kaplan, another Stanford academic and author of “Humans Need Not Apply”, a book that predicts upheaval in the labour market. Gloomiest of all is Martin Ford, a software entrepreneur and the bestselling author of “Rise of the Robots”. He warns of the threat of a “jobless future”, pointing out that most jobs can be broken down into a series of routine tasks, more and more of which can be done by machines.

In previous waves of automation, workers had the option of moving from routine jobs in one industry to routine jobs in another; but now the same “big data” techniques that allow companies to improve their marketing and customer-service operations also give them the raw material to train machine-learning systems to perform the jobs of more and more people. “E-discovery” software can search mountains of legal documents much more quickly than human clerks or paralegals can. Some forms of journalism, such as writing market reports and sports summaries, are also being automated.

Predictions that automation will make humans redundant have been made before, however, going back to the Industrial Revolution, when textile workers, most famously the Luddites, protested that machines and

steam engines would destroy their livelihoods. “Never until now did human invention devise such expedients for dispensing with the labour of the poor,” said a pamphlet at the time. Subsequent outbreaks of concern occurred in the 1920s (“March of the machine makes idle hands”, declared a New York Times headline in 1928), the 1930s (when John Maynard Keynes coined the term “technological unemployment”) and 1940s, when the New York Times referred to the revival of such worries as the renewal of an “old argument”.

As computers began to appear in offices and robots on factory floors, President John F. Kennedy declared that the major domestic challenge of the 1960s was to “maintain full employment at a time when automation…is replacing men”. In 1964 a group of Nobel prizewinners, known as the Ad Hoc Committee on the Triple Revolution, sent President Lyndon Johnson a memo alerting him to the danger of a revolution triggered by “the combination of the computer and the automated self-regulating machine”. This, they said, was leading to a new era of production “which requires progressively less human labour” and threatened to divide society into a skilled elite and an unskilled underclass. The advent of personal computers in the 1980s provoked further hand-wringing over potential job losses.

Yet in the past technology has always ended up creating more jobs than it destroys. That is because of the way automation works in practice, explains David Autor, an economist at the Massachusetts Institute of Technology. Automating a particular task, so that it can be done more quickly or cheaply, increases the demand for human workers to do the other tasks around it that have not been automated.

There are many historical examples of this in weaving, says James Bessen, an economist at the Boston University School of Law. During the Industrial Revolution more and more tasks in the weaving process were automated, prompting workers to focus on the things machines could not do, such as operating a machine, and then tending multiple machines to keep them running smoothly. This caused output to grow explosively. In America during the 19th century the amount of coarse cloth a single weaver could produce in an hour increased by a factor of 50, and the amount of labour required per yard of cloth fell by 98%. This made cloth cheaper and increased demand for it, which in turn created more jobs for weavers: their numbers quadrupled between 1830 and 1900. In other words, technology gradually changed the nature of the weaver’s job, and the skills required to do it, rather than replacing it altogether.

In a more recent example, automated teller machines (ATMs) might have been expected to spell doom for bank tellers by taking over some of their routine tasks, and indeed in America their average number fell from 20 per branch in 1988 to 13 in 2004, Mr Bessen notes. But that reduced the cost of running a bank branch, allowing banks to open more branches in response to customer demand. The number of urban bank branches rose by 43% over the same period, so the total number of employees increased. Rather than destroying jobs, ATMs changed bank employees’ work mix, away from routine tasks and towards things like sales and customer service that machines could not do.

The same pattern can be seen in industry after industry after the introduction of computers, says Mr Bessen: rather than destroying jobs, automation redefines them, and in ways that reduce costs and boost demand. In a recent analysis of the American workforce between 1982 and 2012, he found that employment grew significantly faster in occupations (for example, graphic design) that made more use of computers, as automation sped up one aspect of a job, enabling workers to do the other parts better. The net effect was that more computer-intensive jobs within an industry displaced less computer-intensive ones. Computers thus reallocate rather than displace jobs, requiring workers to learn new skills. This is true of a wide range of occupations, Mr Bessen found, not just in computer-related fields such as software development but also in administrative work, health care and many other areas. Only manufacturing jobs expanded more slowly

than the workforce did over the period of study, but that had more to do with business cycles and offshoring to China than with technology, he says.

So far, the same seems to be true of fields where AI is being deployed. For example, the introduction of software capable of analysing large volumes of legal documents might have been expected to reduce the number of legal clerks and paralegals, who act as human search engines during the “discovery” phase of a case; in fact automation has reduced the cost of discovery and increased demand for it. “Judges are more willing to allow discovery now, because it’s cheaper and easier,” says Mr Bessen. The number of legal clerks in America increased by 1.1% a year between 2000 and 2013. Similarly, the automation of shopping through e-commerce, along with more accurate recommendations, encourages people to buy more and has increased overall employment in retailing. In radiology, says Dr Barani, Enlitic’s technology empowers practitioners, making average ones into experts. Rather than putting them out of work, the technology increases capacity, which may help in the developing world, where there is a shortage of specialists.

And while it is easy to see fields in which automation might do away with the need for human labour, it is less obvious where technology might create new jobs. “We can’t predict what jobs will be created in the future, but it’s always been like that,” says Joel Mokyr, an economic historian at Northwestern University. Imagine trying to tell someone a century ago that her great-grandchildren would be video-game designers or cybersecurity specialists, he suggests. “These are jobs that nobody in the past would have predicted.”

Similarly, just as people worry about the potential impact of self-driving vehicles today, a century ago there was much concern about the impact of the switch from horses to cars, notes Mr Autor. Horse-related jobs declined, but entirely new jobs were created in the motel and fast-food industries that arose to serve motorists and truck drivers. As those industries decline, new ones will emerge. Self-driving vehicles will give people more time to consume goods and services, increasing demand elsewhere in the economy; and autonomous vehicles might greatly expand demand for products (such as food) delivered locally.

Only humans need apply

There will also be some new jobs created in the field of AI itself. Self-driving vehicles may need remote operators to cope with emergencies, or ride-along concierges who knock on doors and manhandle packages. Corporate chatbot and customer-service AIs will need to be built and trained and have dialogue written for them (AI firms are said to be busy hiring poets); they will have to be constantly updated and maintained, just as websites are today. And no matter how advanced artificial intelligence becomes, some jobs are always likely to be better done by humans, notably those involving empathy or social interaction. Doctors, therapists, hairdressers and personal trainers fall into that category. An analysis of the British workforce by Deloitte, a consultancy, highlighted a profound shift over the past two decades towards “caring” jobs: the number of nursing assistants increased by 909%, teaching assistants by 580% and careworkers by 168%.

Focusing only on what is lost misses “a central economic mechanism by which automation affects the demand for labour”, notes Mr Autor: that it raises the value of the tasks that can be done only by humans. Ultimately, he says, those worried that automation will cause mass unemployment are succumbing to what economists call the “lump of labour” fallacy. “This notion that there’s only a finite amount of work to do, and therefore that if you automate some of it there’s less for people to do, is just totally wrong,” he says. Those sounding warnings about technological unemployment “basically ignore the issue of the economic response to automation”, says Mr Bessen.

But couldn’t this time be different? As Mr Ford points out in “Rise of the Robots”, the impact of automation this time around is broader-based: not every industry was affected two centuries ago, but every industry uses computers today. During previous waves of automation, he argues, workers could switch from one kind of routine work to another; but this time many workers will have to switch from routine, unskilled jobs to non-routine, skilled jobs to stay ahead of automation. That makes it more important than ever to help workers acquire new skills quickly. But so far, says Mr Autor, there is “zero evidence” that AI

is having a new and significantly different impact on employment. And while everyone worries about AI, says Mr Mokyr, far more labour is being replaced by cheap workers overseas.

Another difference is that whereas the shift from agriculture to industry typically took decades, software can be deployed much more rapidly. Google can invent something like Smart Reply and have millions of people using it just a few months later. Even so, most firms tend to implement new technology more slowly, not least for non-technological reasons. Enlitic and other companies developing AI for use in medicine, for example, must grapple with complex regulations and a fragmented marketplace, particularly in America (which is why many startups are testing their technology elsewhere). It takes time for processes to change, standards to emerge and people to learn new skills. “The distinction between invention and implementation is critical, and too often ignored,” observes Mr Bessen.

What of the worry that new, high-tech industries are less labour-intensive than earlier ones? Mr Frey cites a paper he co-wrote last year showing that only 0.5% of American workers are employed in industries that have emerged since 2000. “Technology might create fewer and fewer jobs, while exposing a growing share of them to automation,” he says. An oft-cited example is that of Instagram, a photo-sharing app. When it was bought by Facebook in 2012 for $1 billion, it had tens of millions of users, but only 13 employees. Kodak, which once employed 145,000 people making photographic products, went into bankruptcy at around the same time. But such comparisons are misleading, says Marc Andreessen. It was smartphones, not Instagram, that undermined Kodak, and far more people are employed by the smartphone industry and its surrounding ecosystems than ever worked for Kodak or the traditional photography industry.

Is this time different?

So who is right: the pessimists (many of them techie types), who say this time is different and machines really will take all the jobs, or the optimists (mostly economists and historians), who insist that in the end technology always creates more jobs than it destroys? The truth probably lies somewhere in between. AI will not cause mass unemployment, but it will speed up the existing trend of computer-related automation, disrupting labour markets just as technological change has done before, and requiring workers to learn new skills more quickly than in the past. Mr Bessen predicts a “difficult transition” rather than a “sharp break with history”. But despite the wide range of views expressed, pretty much everyone agrees on the prescription: that companies and governments will need to make it easier for workers to acquire new skills and switch jobs as needed. That would provide the best defence in the event that the pessimists are right and the impact of artificial intelligence proves to be more rapid and more dramatic than the optimists expect.

NYT

New Tools Needed to Track Technology’s Impact on Jobs, Panel SaysBy STEVE LOHRAPRIL 13, 2017

Automation fears have stirred many fanciful visions of the future. An expert panel calls for new tools of analysis to help workers adapt. Credit Kristian Hammerstad

America needs new tools for the timely measurement and monitoring of technology, jobs and skills to cope with the advance of artificial intelligence and automation, an expert panel composed mainly of economists and computer scientists said in a new report.

The panel’s recommendations include the development of an A.I. index, analogous to the Consumer Price Index, to track the pace and spread of artificial intelligence technology. That technical assessment, they said, could then be combined with detailed data on skills and tasks involved in various occupations to guide education and job-training programs.

A public-private collaboration, they added, is necessary to create such tools because information from many sources will be the essential ingredient. Those information sources range from traditional government statistics to the vast pools of new data from online services like LinkedIn and Udacity that can be tapped to gain insights on skills, job openings and the effectiveness of training programs.

“We’re flying blind into this dramatic set of economic changes,” Erik Brynjolfsson, an economist at the Massachusetts Institute of Technology’s Sloan School of Management, said in an interview.

Mr. Brynjolfsson was a co-chairman of the 13-member panel that drafted the 184-page report , which was published on Thursday by the National Academies of Sciences, Engineering and Medicine, a nonprofit organization whose studies are intended as objective analysis to inform public policy. He and the panel’s other co-chairman, Tom Mitchell, a computer scientist at Carnegie Mellon University, also wrote a separate commentary in the journal Nature that was published on Thursday, explaining the problem.

Both the report and commentary were spurred by the advances in A.I. in recent years, including document-reading software and self-driving cars, which promise to make inroads into work done by humans. That prospect has created angst for many American workers about the difficulties of adapting to technological change and the failure of institutions to help them.

Yet technologists and academics still differ sharply on how fast the next wave of automation will proceed and how many occupations will be affected. That prompted the panelists to suggest the new data-monitoring tools and the pulling together of government and online data sources to sort through the consequences.

Photo

An autoworker feeds aluminum panels to robots at a Ford plant in Kansas City, Mo. Credit Dave Kaup/Reuters

Those moves could eventually give a worker in a declining occupation useful information about a more promising occupation, with some similar skills but also requiring some new ones, Mr. Mitchell said. Then the software tool might also pull information on job placement rates for courses that teach those new skills.

That style of data-driven decision-making is a hallmark of internet companies like Amazon and Google, and it has been increasingly embraced across corporate America. “There’s no reason government can’t do that,” Mr. Brynjolfsson said.

In recent years, the federal government has made considerable progress in integrating its surveys of businesses and households with other data it collects, including information on foreign trade, payrolls and unemployment, said John Haltiwanger, a professor at the University of Maryland and former chief economist of the Census Bureau.

“But our surveys are not really designed to track technology or its impact,” said Professor Haltiwanger, who was a member of the expert panel. “The best shot at that is the private sector data.”

A broad national initiative, perhaps with the Bureau of Labor Statistics and the Bureau of Economic Analysis setting rules for private sector data sharing and privacy protections, might not be possible, Mr. Brynjolfsson conceded. But he and Mr. Mitchell wrote in Nature, “Perfection here is not a prerequisite for utility — anything is better than flying blind.”

One program that embodies the panel’s recommendations is Skillful, a collaboration of the Markle Foundation, LinkedIn, Arizona State University and edX, a nonprofit provider of online courses. The partners are working with employers, educators and local governments in Colorado and the Phoenix metropolitan area to link jobs, skills and training more tightly.

Past times of economic turmoil have led to new kinds of economic data collection and analysis.

At the outset of the Depression in 1929, for example, there was no measure of national economic activity or reliable information on unemployment. In June 1930, based on scattered reports of improvement, President Herbert Hoover prematurely declared, “The Depression is over.”

To address the information gap of its day, the government hired economists and statisticians to come up with a scientific method for measuring the national economy. In 1934, a team led by Simon Kuznets

published its report on how to calculate national income. The field of econometrics took a big step forward, and policy makers were less in the dark.

The time has now come, the expert panel suggested, for a similar effort to adapt to the modern digital economy.