“They’re always polite, they always upsell, they never take a vacation, they never show up late, there’s never a slip-and-fall, or an age, sex, or race discrimination case,” said Andrew Puzder, then chief executive of Hardee’s Food Systems Inc., a restaurant chain headquartered in Tennessee. He was talking about swapping employees for machines.
Statements like these give workers reason to worry*. The advent of a jobless economy raises concern because tasks traditionally performed by humans are being—or are at risk of being—taken over by robots, especially those enabled with artificial intelligence. The number of robots operating worldwide is rising quickly. By 2019, 1.4 million new industrial robots will be in operation, raising the total to 2.6 million worldwide. Robot density per worker in 2018 is the highest in Germany, Korea, and Singapore. Yet in all of these countries, despite the high prevalence of robots, the employment rate remains high.
The Robot Takes the Job from the Young Worker
Young workers may be more affected by automation than older workers. Although the adoption of robots did not have any substantial net effect on employment in Germany, it reduced the hiring of young entrants. For this reason, the effects of automation can be different in countries that are aging compared with those that have young populations and anticipate large numbers of new labor market entrants.
Yes, robots are replacing workers, but it is far from clear to what extent. Overall, technological change that replaces routine work is estimated to have created more than 23 million jobs across Europe from 1999 to 2016, or almost half of the total increase in employment over the same period. Recent evidence for European countries suggests that although technology may be replacing workers in some jobs, overall it raises the demand for labor. For example, instead of hiring traditional loan officers, JD Finance, a leading fintech platform in China, created more than 3,000 risk management or data analysis jobs to sharpen algorithms for digitized lending. Technological progress leads to the direct creation of jobs in the technology sector. People are increasingly using smartphones, tablets, and other portable electronic devices to work, organize their finances, secure and heat their homes, and have fun. Workers create the online interfaces that drive this growth. With consumer interests changing fast, there are more opportunities for people to pursue careers in mobile app development and virtual reality design.
Technology Creates Jobs
Technology has also facilitated the creation of jobs through working online or joining the so-called gig economy. Andela, a U.S. company that specializes in training software developers, has built its business model on the digitization of Africa. It has trained 20,000 software programmers across Africa using free online learning tools. Once qualified, programmers work with Andela directly or join other Andela clients across the world. The company aims to train 100,000 African software developers by 2024. Ninety percent of its workers are in Lagos, Nigeria, with other sites in Nairobi, Kenya, as well as Kampala, Uganda.
Technology increases proximity to markets, facilitating the creation of new, efficient value chains. In Ghana, Farmerline is an online platform that communicates with a network of more than 200,000 farmers in their native languages via mobile phone. It provides information on the weather and market prices, while collecting data for buyers, governments, and development partners. The company is expanding to include credit services.
During this process of technology adoption some workers will be replaced by technology. Workers involved in routine tasks that are “codifiable” are the most vulnerable. The examples are numerous. More than two-thirds of robots are employed in the automotive, electrical/electronics, and metal and machinery industries. Based in China, Foxconn Technology Group, the world’s largest electronics assembler, cut its workforce by 30 percent when it introduced robots into the production process. When robots are cheaper than the existing manufacturing processes, firms become more amenable to relocating production closer to consumer markets. In 2017 3-D printing technologies enabled the German company Adidas to establish two “speed factories” for shoe production: one in Ansbach, Germany, and the other in Atlanta in the United States, eliminating more than 1,000 jobs in Vietnam. In 2012 the Dutch multinational technology company Philips Electronics shifted production from China back to the Netherlands.
Many Service Jobs Will be Automatized
Some service jobs are also vulnerable to automation. Mobileye of Israel is developing driverless vehicle navigation units. Baidu, the Chinese technology giant, is working with King Long Motor Group, China, to introduce autonomous buses in industrial parks. Financial analysts, who spend much of their time conducting formula-based research, are also experiencing job cuts: Sberbank, the largest bank in the Russian Federation, relies on artificial intelligence to make 35 percent of its loan decisions, and it anticipates raising that rate to 70 percent in less than five years. “Robot lawyers” have already replaced 3,000 human employees in Sberbank’s legal department. The number of back-office employees will shrink to 1,000 by 2021, down from 59,000 in 2011. Ant Financial, a fintech firm in China, uses big data to assess loan agreements instead of hiring thousands of loan officers or lawyers.
Nevertheless, it is impossible to put a figure on the level of job displacement that will take place overall. Even the most well-known economists have experienced little success with this exercise. In 1930 John Maynard Keynes declared that technology would usher in an age of leisure and abundance within a hundred years. He mused that everyone would have to do some work if they were to be content, but that three hours a day would be quite enough. The world in 2018 is far from this kind of reality.
How Many Jobs are Affected?
Although quantifying the impact of technological progress on job losses continues to challenge economists, estimates abound. Those estimates vary widely (see figure). For Bolivia, job automation estimates range from 2 to 41 percent. In other words, anywhere from 100,000 to 2 million Bolivian jobs may be automated in 2018. The range is even wider for advanced economies. In Lithuania, from 5 to 56 percent of jobs are at risk of being automated. In Japan, from 6 to 55 percent of jobs are thought to be at risk.
FIGURE Estimates of the percentage of jobs at risk from automation vary widely
Sources: WDR 2019 team, based on World Bank (2016)
Note: The figures represent the highest and lowest estimates of the percentage of jobs at risk of automation in economies for which more than one estimate has been produced by different studies. A job is at risk if its probability of being automated is greater than 0.7.
The wide range of predictions illustrates the difficulty of estimating technology’s impact on jobs. Most estimates rely on automation probabilities developed by machine learning experts at the University of Oxford. The experts were asked to categorize a sample of 70 occupations taken from the O*NET online job database used by the U.S. Department of Labor as either strictly automatable or not (1–0). Relying on these probabilities, initial estimates placed 47 percent of U.S. occupations at risk of automation. Basing probabilities on the opinion of experts is instructive but not definitive. Moreover, using one country’s occupational categories to estimate possible job losses from automation elsewhere is problematic.
Will We See the Same Development in All Countries?
Job loss predictions do not accurately incorporate technology absorption rates, which are often painstakingly slow and differ not only between countries but also across firms within countries. The absorption rate therefore affects the potential for technology to destroy jobs. The use of mobile telephony, for example, spread faster than earlier technologies, but the Internet has been comparatively slow to take hold in many cases, particularly among firms in the informal sector. The uptake of mechanization in agriculture presents a similar picture. Persistent trade barriers, the relatively low cost of labor compared with that of agricultural machinery, and poor information all contribute to the low rates of mechanization in low-income and some middle-income countries. Even for the textile industry’s spinning jenny, the relatively low cost of labor delayed its introduction in France and India in 1790 France had only 900 spinning jennies compared with 20,000 in Great Britain. The prevalence of automation versus labor continues to vary across and within countries, depending on the context.
The Picture on top: Kzenon
*The article is from Chapter 1: THE CHANGING NATURE OF WORK. WORLD DEVELOPMENT REPORT 2019. The World Bank 2019.
Categories: Development, Future Work, Technology, World