How Is New Technology Changing Job Design?

4m ago
1 Views
0 Downloads
354.01 KB
10 Pages
Last View : 11d ago
Last Download : n/a
Upload by : Hayden Brunner
Transcription

MICHAEL GIBBSUniversity of Chicago, USA, and IZA, GermanyHow is new technology changing job design?Machines’ ability to perform cognitive, physical, and social tasksis accelerating, dramatically changing jobs and labor marketsKeywords: job design, technology, artificial intelligence, cognitive tasks, labor market polarizationELEVATOR PITCHThe information technology revolution has had dramaticeffects on jobs and the labor market. Many routine andmanual tasks have been automated, replacing workers.By contrast, new technologies complement non-routine,cognitive, and social tasks, making work in such tasks moreproductive. These effects have polarized labor markets:While low-skill jobs have stagnated, there are fewer andlower paid jobs for middle-skill workers, and higher pay forhigh-skill workers, increasing wage inequality. Advances inartificial intelligence may be accelerating computers’ abilityto perform cognitive tasks, heightening concerns aboutautomation of even high-skill jobs.Share of workers at high risk ( 70%) of automationwithin 10 to 20 elandDenmarkFranceUSAll countriesCanadaItalyNetherlandsCzech e: [1].KEY FINDINGSProsTechnology complements many tasks, increasingproductivity, quality, and innovation.Big data and machine learning are increasingmachines’ ability to perform cognitive, physical,and even some social (language) tasks.Greater access to data, analysis tools, andtelecommunications allows many workers tofocus more on social interactions, collaboration,continuous improvement, and innovation.Technology makes many high-skill jobs moreintrinsically motivating, enabling more tasks,skills, and decentralization.ConsMachines substitute for humans in many manualand routine jobs.The pace at which machines gain the ability toperform cognitive tasks is faster than in the past,making adaptation by workers more difficult.Labor markets have polarized and inequality hasrisen, with relatively less demand for mid-skillworkers and increased value for high-skill workers.Technology makes many middle-skill jobs lessintrinsically motivating, with fewer tasks and skills,and more centralization and monitoring.AUTHOR’S MAIN MESSAGETechnology has opposing effects on jobs. It facilitates automation, creating fewer and less motivating middle-skilljobs. Conversely, it complements social and innovation tasks, creating more interesting low- and high-skill jobs. Thiscauses labor market polarization, “hollowing out” demand for middle-skill jobs, and increasing wage inequality.Some claim that computers will soon replace many workers as artificial intelligence advances. Others are skeptical, asprevious technological advances did not lead to mass unemployment. Policymakers should encourage technology thatcomplements employees’ work, and should foster education and training that help workers adapt to change.How is new technology changing job design? IZA World of Labor 2017: 344doi: 10.15185/izawol.344 Michael Gibbs March 2017 wol.iza.org1

MICHAEL GIBBS How is new technology changing job design?MOTIVATIONSince at least the Luddite movement of 1811–1816, the effects of new technology on jobsand employment have generated great controversy. In the last few decades, enormousimprovements in information and telecommunications technologies (ICT) have haddramatic effects, benefiting some workers, but eliminating others’ jobs.The current debate focuses on two issues: How far, and how fast, will job automationproceed? Recent advances in artificial intelligence methods are pushing past previous limitson the types of tasks that can be automated. Machines increasingly perform cognitivetasks, use natural language, and have greater dexterity and mobility. Some observersclaim a high fraction of jobs are at risk of being automated, including high-skill jobs forthe first time, potentially leading to large-scale unemployment. Others are more skeptical,noting barriers to technological advancement and implementation, and pointing out thatlabor markets have always managed to absorb new technologies in the past.DISCUSSION OF PROS AND CONSHow does technological change affect job design? Think of a job as a set of tasks thatrequire various types of employee skills. New technology raises relative employee produ ctivity in some tasks, and replaces employees in other tasks. Firms respond by changingjob design—the mix of tasks assigned to workers—and subsequently their demand forworkers with different skills.Early technology tended to increase productivity of low-skill manual laborers by providingbetter tools, machinery, and cheaper raw materials. This was reflected in gradualmechanization of agriculture, and the movement from artisan to factory manufacturingin the late 1800s [2]. However, by about 1910, new technology began favoring middleand high-skill workers. Factories shifted to electric power, which facilitated batch orcontinuous production methods, and assembly lines. Factory foremen, machinists, andmanagers became more productive, overseeing more resources and output. Meanwhile,many manual jobs were mechanized.This is an early example of a general point. Technology sometimes complements employees byincreasing their ability to perform certain tasks, and sometimes substitutes for employees byautomating some or all of their tasks. It thus changes job design by refocusing the employeeon tasks that are difficult to automate, and eliminating tasks that are easy to automate.Additionally, the effect of new technology can change over time. Initially, it complementedlow-skill work. Later, it substituted for it while complementing middle- and high-skill work.Today it complements high-skill work but often substitutes for middle-skill work. It isreasonable to expect that ICT’s effects may change again in the future.Automating tasks (in machinery or software) has several advantages. It reduces variationsince machines tend to perform identically every time. This lowers uncertainty andhelps improve the quality of decisions, products, or services. Machines, and particularlycomputers, often generate large economies of scale. Firms can avoid the complexities ofmanaging employees, including conflict, incentive problems, and absenteeism. Therefore,if the cost of automating a task falls far enough, firms are likely to automate that task.Which tasks are easiest to automate? Those that are most easily understood, optimized,and codified in advance. Thus, routine, simple tasks have been most susceptible toIZA World of Labor March 2017 wol.iza.org2

MICHAEL GIBBS How is new technology changing job design?mechanization and computerization [3]. As noted above, initially, automation was ofmanual tasks in manufacturing. Experts such as Frederick Taylor devised methods tobreak production into specific steps, and then optimize each step. Doing so codifiedthe task, which facilitated mechanization. From the 1970s onward, the ICT revolutionenabled similar automation of many routine, predictable tasks in clerical and white-collarjobs. Work involving information processing, producing financial forms, making routinecalculations, etc., was easily taken over by computers. This “re-engineering” eliminatedmany middle-skill jobs (e.g. clerical work, data entry, book-keeping), and reduced thenumber of layers in corporate hierarchies.Simpler, more stable, and predictable environments favor automation for two reasons:ease of optimization and technological longevity [4]. For tasks to be automated, the firmmust invest resources in analyzing and optimizing that part of the process. Perfecting partof a process takes resources (e.g. consultants, total quality management methods). Thisinvestment will be more profitable if the optimization problem is easier, as is the case withsimpler products and product lines. It will also be more profitable if the new knowledge canbe deployed longer in the future, as is the case with stable and predictable environments.For example, UPS (a worldwide package delivery company) famously optimized the jobof delivery truck drivers, even to the extent of teaching them how to step into the truck inthe quickest possible way. Its business was very simple (deliver a package from one placeto another), as well as stable and predictable (methods evolved little over more than 100years, from bicycles to motorcycles to trucks, then to airplanes for long distances).Which tasks are more difficult to automate? First, not all manual tasks have proven easyto automate. Physical tasks sometimes involve fine motor coordination and dexterity,which machines have not been able to replicate. They also often involve observing andinterpreting the worker’s physical environment, as well as moving within random physicalspaces. Computers and machines have historically lacked these capabilities, includingvision and image recognition (Figure 1).Cognitive tasks have also been difficult to automate. They require higher-order thinkingskills, while computers have tended to only perform specific, programmed operations.Instead of being automated, jobs involving analysis, decision making, abstract thinking,learning, innovation, and creativity are often complemented by new technology. Forexample, the job of an aircraft design engineer has changed dramatically. In the past, itFigure 1. Types of tasks most difficult to automateType of taskAttributes that are difficult toautomateExampleNon-routineComplex; unpredictable; changingTax law; law enforcement; computer networkingManualObject recognition; mobility inunmapped space; fine dexteritySorting random objects; restaurant table service;surgeryCognitiveManaging change; continuousimprovement; creativity; innovation;abstract analysisOrganizational restructuring; total quality management;art; pharmaceutical research; economic theorySocialService; negotiation; teaching;collaboration; management/coordination;leadershipNursing; sales; professor; orchestra performer; projectmanagement; CEOSource: Author’s own composition.IZA World of Labor March 2017 wol.iza.org3

MICHAEL GIBBS How is new technology changing job design?involved substantial tedious work, producing complex blueprints by hand calculation anddrawing. Now engineers have computers that perform these tasks, freeing them up tofocus more on design and complex configuration options [5].Social tasks have also proven difficult to automate. Computers and robots do not havethe ability to empathize with colleagues and customers, inspire employees, use intuition,or listen and communicate with subtlety. Tasks involving social interactions, often in lowskill service jobs and high-skill management jobs, have largely avoided automation. Socialskills have become increasingly valuable in the labor market, and employment growth hasbeen largest in jobs that are high in both cognitive and social skill requirements [6]. Thatis, social and cognitive skills appear to be complementary.Summing up, a job is a bundle of manual, cognitive, and social tasks. New technology allowsfirms to automate some tasks, taking them from workers and performing them insteadwith machines and computers. It also allows firms to provide workers with information,data, analysis, and communication tools that increase their ability to perform other tasks.Thus, the effect of technology on job design rests on a substitute–complement continuum.For some jobs, most or all tasks can be automated. For some jobs, few tasks can beautomated, but many can be complemented by technology. Other jobs lie in between,with some tasks automated, some unaffected, and some complemented.For example, some medical diagnostic tests have been automated, eliminating manymedical technician jobs. Some nursing tasks have been replaced by bedside machinesthat monitor patients and dispense medicine, but the nurse’s interaction with the patientis largely impossible to automate. Finally, virtually all surgeries are still performed byhumans, but surgeons have advanced tools that allow them to perform these surgeriesmore quickly, safely, and effectively.This process can lead to dramatic differences in employees’ work [7]. For jobs that aremostly automated, managers tend to make most or all decisions and workers simplyperform their prescribed tasks. This is because much of the process has already beenoptimized, so the worker can add little new knowledge to the job, and few decisionsor changes need to be made. These jobs usually require few skills, involve only a fewrepetitive tasks, require little thinking by the worker, and therefore tend to have lowintrinsic motivation. By contrast, jobs that are complemented by technology tendto require more skills, including problem-solving and social skills. They tend to makemore use of decentralization so that employees learn, and then develop, test, andimplement ideas and solutions [4]. As a result, such jobs have high intrinsic motivation[8]. Consistent with these ideas, investment in ICT and research and development arepositively associated with more enriched job designs, large-scale organizational change,continuous improvement, and greater competition.Other effects of technological changeTechnology is changing the employer–employee relationship, and even what it means to bean “employee”; however, such effects are only briefly touched upon here as they are not themain scope of this article. It is now easy to collaborate remotely by file and data sharing,email, and videoconference. Except when joint work needs to be synchronous and face-toface, this can change traditional modes of work, employment, and firm structures. TheseIZA World of Labor March 2017 wol.iza.org4

MICHAEL GIBBS How is new technology changing job design?developments facilitate the globalization of firms, but also enable the outsourcing andoffshoring of jobs beyond firm boundaries. They enable new arrangements, with moreflexibility in tasks, total hours, timing, and location of work. Recently, this has manifestedin the “gig economy,” in which an increasing fraction of the labor force is employed inshort-term, part-time arrangements without attachment to a traditional employer. Thenet effect of these changes on workers and firms are not well understood; society shouldexpect to see further evolution in these areas.ICT may further change how firms motivate employees. On the one hand, many jobs havebecome more decentralized. Employees with greater discretion are usually given strongerpay for performance in order to align their goals and decisions with firm objectives. Onthe other hand, technology leads to greater centralization in some other jobs, and makesit possible to monitor and assess employees in new ways. For example, identificationbadges can be designed to track an employee’s location, note which colleagues they spendtime with, monitor the tone of employee conversations, and note how they stand relativeto each other during those conversations. Machine-learning algorithms can then, forinstance, analyze such data to evaluate employee “leadership potential.”Labor market polarizationAs new technology substitutes or complements different types of tasks, it changes therelative demand for skills needed to perform those tasks. Skills associated with tasks thatmachines can now perform tend to see a relative decline in demand, while those associatedwith tasks that are complemented by new technology see a relative rise in demand. Thus,technological change affects the relative compensation of workers with different skilltypes. The supply of workers with different skill types will also change. However, laborsupply tends to change slowly since it requires changes in education and training. Thismeans that skills and wages tend to be highly correlated. For that reason, labor economicsresearchers often proxy “skills” by the level of pay.Automation in modern times has tended to focus on middle-skill jobs. High-skill jobscomprise cognitive tasks, social skills (management and leadership), and creativity. Whilesome low-skill jobs have been automated, those requiring greater dexterity, teamwork, orinteractions with customers have not been widely automated. By contrast, middle-skilljobs tend to involve routine information processing, calculation, and decision making.They have therefore been hardest hit by automation with the advent of cheap, powerfulcomputers, and greater access to data.This pattern is often termed labor market polarization [9]. Polarization has two aspects.First, the relative share of low- and high-skill jobs has increased, with a “hollowing out”of the share of middle-skill jobs. Second, this has increased wage inequality, since middleskill jobs have fallen in prominence, while complementarity with technology has increasedrelative compensation for high-skilled workers; Figure 2 illustrates this for OECD countries.Polarization is a relatively new phenomenon. Until recently, labor markets reflected skillbiased technological change in which technology favored workers with more skill relativeto those with less. What is different now is that computers can perform analysis and, tosome extent, cognitive tasks; hence, in the last three decades, routine-biased technologicalchange has emerged.IZA World of Labor March 2017 wol.iza.org5

MICHAEL GIBBS How is new technology changing job design?Figure 2. Change in occupational employment shares: Low-, middle-, and high-wageoccupations, 1993–201015%Low payingMiddle payingHigh paying12%9%6%3%0%lgar 8.5–7.6–6.7–9.6–12.0–14.9–18%Source: Autor, D. H. “Why are there still so many jobs? The history and future of workplace automation.” Journal ofEconomic Perspectives 29:3 (2015): 3–30 [9].How far and how fast is automation of tasks proceeding?How technology affects job design has recently changed. Initially, computers hadlargely automated tasks that could be well-defined and guided by humans, either viatraditional computer programs that specify what the computer should do, or expertsystems designed to categorize and replicate human decision making. Recently, however,computer scientists have made strides in machine learning, in which computers develop,evaluate, and refine their own algorithms, with little or no human intervention. Thispresents a new approach: automation of cognitive tasks. Moreover, such algorithms haveimproved mobility, dexterity, vision, and object recognition in robotics. For example, oneinfluential paper from 2003 describes deciphering the signature on a check and drivinga car as tasks that were difficult to automate [3]. However, checks are now routinelyprocessed by computers that decipher handwriting, and driverless cars are being testedin several cities.The development of computers that can learn is a potentially dramatic change in taskautomation. How far these developments are likely to proceed, and how quickly, is thesubject of great debate. Some argue that the pace of automation has accelerated, includingfor the first time in high-skilled jobs [10]. One study analyzed the task content of 7,000jobs and concluded that nearly half, including many high-skilled jobs, are at high risk ofautomation in the next 10 to 20 years [11]. If that prediction proves true, the implicationsfor labor markets could be dramatic.However, the study provoked controversy. For example, it analyzed risk of automation atthe job level, but jobs comprise a set of tasks, some of which might be automated, whileothers might not. The illustration on page 1 comes from a work that refined the previousstudy; the authors concluded that the fraction of jobs at high risk for automation is notIZA World of Labor March 2017 wol.iza.org6

MICHAEL GIBBS How is new technology changing job design?Machine learningComputers were initially deployed in the workplace for routinized information-processingtasks, since such tasks could be programmed. This was very similar to mechanization of routinetasks during the Industrial Revolution. Recently, the role of computers at work has begun tochange: they can now learn and evolve from experience. This is largely due to the arrival ofbig data—massive increases in affordable computer processing power, storage capacity, andinformation (numeric and otherwise)—as the new methods are extremely data-intensive.A set of tools has been used to develop computers that perform some cognitive, naturallanguage, and image recognition tasks. These tools also improve robots’ mobility anddexterity. The methods developed as a result of recent advances and can analyze manytypes of information, such as graphical images, language, and maps. Unlike traditionalstatistics, they usually do not attempt to fit a pre-specified model, but are designed todiscover complex relationships between different pieces of information. They are provingextremely flexible and widely applicable.Data mining: exploratory techniques to uncover patterns in data.Machine learning: similar to data mining, but directed at specific goals. This sub-field ofartificial intelligence develops algorithms for prediction and/or decision making. Thecomputer is “trained” with example inputs and outputs (e.g. signatures and names onchecks), and iteratively develops an algorithm to perform a desired task on any new data.The technique is closely related to statistics, since it involves iteratively fitting a model todata to minimize a cost function.Neural networks: branch of machine learning inspired by neuroscience. Iteratively developsan algorithm in which a network of artificial neurons process and pass data to each other,sometimes in complicated configurations. This method is especially effective for very complexproblems, since it can handle millions of dimensions, and relationships between objects (e.g.data or symbols) can be non-linear and “tangled” (akin to endogeneity in statistics).Deep learning: somewhat ill-defined term for advanced techniques in which multiple neuralnetworks work together in layered (hierarchical) fashion. Designed to model high-levelabstractions generating the data. For example, one stage might model how to effectivelyrepresent an image (a set of pixels or a set of edges between colors). The next stagewould use that output to determine some property of the image, such as whether it is ahuman face. Other steps might then follow. This method has proven particularly usefulfor mimicking vision and natural language.50%, but closer to 5–10% [1]. Jobs at least risk of automation are estimated to involvegreater use of deductive reasoning, originality, communication, training, problem solving,and reading and writing. They also have greater requirements for pre-job education ortraining [1], [12]. However, considerably more work needs to be done before researcherswill be confident answering these questions.LIMITATIONS AND GAPSA significant limitation of the current debate on technological change is that it is difficultto predict future advancements, effects on job design, and labor market responses.Computer scientists are uncertain of how much progress will be made, and at what pace.Artificial intelligence has proceeded in spurts, with occasional advancements followed byIZA World of Labor March 2017 wol.iza.org7

MICHAEL GIBBS How is new technology changing job design?slower periods in which obstacles have proven difficult to overcome. Furthermore, expertsmay overestimate likely progress in their own field.Even if one knew the extent of future change, research on potential task automation isspeculative and at an early stage. Equally important, but understudied, is the likelihoodthat ICT and machine learning might further complement tasks rather than automatethem. Mechanisms by which technology complements work are not as well understoodas those by which it substitutes. More evidence is needed on how machine learning andother technologies are implemented, and on how they substitute or complement differenttasks, as well as the ultimate effect on job designs.The speed at which new technology will be implemented is uncertain. Past experience,including the personal computer revolution of the 1980s, suggests adoption can be slowand difficult. It takes time for organizations to learn practical implementation. Changeis slow, complex, and may require high pressure to succeed. Indeed, research indicatesthat changes in jobs to exploit routine-biased technical change is more significant duringrecessions. Finally, new technology often faces regulatory hurdles, as well as resistancefrom political and special interest groups.The future extent of labor market polarization is likewise unclear. New technology has notalways complemented high-skill jobs, and may not in the future. Labor market effects ofnew technology will depend on complex interactions between skill demand and supply,on how technology is deployed, as well as on trade. Job design and work-focused researchand design are endogenous. As middle-skill workers become relatively cheap, firms aremotivated to find ways in which technology can complement their work. For example,health care providers are investing in technology that would allow nurses, home healthaides, and other middle-skill workers (who are relatively inexpensive compared to doctors)to perform some diagnostics and provide limited patient treatment. Some developmentsare likely to lead students, employees, and firms to change the type of skills they investin. International trade allows some types of tasks to be offshored more than others,affecting relative demand for different types of skills [13]. Trade also affects job design,as ICT eliminates geographical barriers to collaboration or offshoring. These interactionsare not yet fully understood, and are likely to change in the future.SUMMARY AND POLICY ADVICESome believe the pace of automation, including of high-skill cognitive and social tasks, isaccelerating. However, one should not overreact. New technology has always generateddire labor market predictions that have never come to fruition. Technological change hasnot always complemented high-skill work while replacing low-skill jobs, and may not inthe future either. Humans are capable of much that is hard to automate, even with theadvent of artificial intelligence. Technology can be used to help people focus on customerservice, artisanal and craft work, innovation, education, and more. Furthermore, ICTimproves productivity and quality, and generates new products and services. Thesegenerate growth, which can improve labor demand [13].Much of the research on robotics and artificial intelligence is aimed at mimicking humans,which biases toward automation. Policymakers should encourage research into howtechnology can instead augment human creativity and collaboration, particularly inmiddle- and low-skill jobs.IZA World of Labor March 2017 wol.iza.org8

MICHAEL GIBBS How is new technology changing job design?What skills are most likely to be valuable with future technological change? First, abstractthinking, analytical, and problem solving skills. For this reason, mathematics, statistics,science, engineering, and economics have risen in prominence. Second, creativity, andsocial and communication skills. Educational institutions need to teach this combinationof skills.Many laws and regulations were established when unions and long careers in a singlelarge corporation were the norm. These often impose detailed, inflexible arrangementsthat may no longer fit the current and future labor market. Policies that loosen the formaldefinition of an “employee” would enable people and firms to more flexibly and creativelyrealize ICT’s potential to enrich careers and improve work–life balance. They wouldincrease productivity and should encourage innovation.AcknowledgmentsThe author thanks two anonymous referees and the IZA World of Labor editors for manyhelpful suggestions on earlier drafts. He also thanks Alec Levenson and Cindy Zoghi fordiscussions on this topic during their collaboration, Geoffrey Gibbs and Kathryn Ierulli forvaluable comments, and Hieu Nguyen for research assistance. Financial support from theUniversity of Chicago Booth School of Business is gratefully acknowledged.Competing interestsThe IZA World of Labor project is committed to the IZA Guiding Principles of Research Integrity.The author declares to have observed these principles. Michael GibbsIZA World of Labor March 2017 wol.iza.org9

MICHAEL GIBBS How is new technology changing job design?REFERENCESFurther readingBrynjolfsson, E., and A. McAfee. Race Against the Machine. Lexington, MA: Digital Frontier Press, 2011.Domingos, P. The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World.New York: Basic Books, 2015.Key references[1]Arntz, M., T. Gregory, and U. Zierahn. The Risk of Automation for Jobs in OECD Countries: AComparative Analysis. OECD Social, Employment and Migration Working Papers No. 189, 2016.[2]Goldin, C., and L. F. Katz. “The origins of technology–skill complementarity.” Quarterly Journal ofEconomics 113:3 (1998): 693–732.[3]Autor, D. H., F. Levy, and R. J. Murnane. “The skill content of recent technological change: Anempirical exploration.” Quarterly Journal of Economics 118:4 (2003): 1279–1333.[4]Lindbeck, A., and D. Snower. “Multitask learning and the reorganization of work: FromTayloristic to holistic organization.” Journal of Labor Economics 18:3 (2000): 353–376.[5]Levy, F., and R. J. Murnane. The New Division of Labor: How Computers are Creating the Next JobMarket. Princeton, NJ: Princeton University Press, 2005.[6]Deming, D. J. The Growing Importance of Social Skills in the Labor Market. Harvard UniversityWorking Paper, 2016.[7]Bresnahan, T., E. Brynjolfsson, and L. M. Hitt. “Information technology, workplaceorganization, and the demand for skill

Technology has opposing effects on jobs. It facilitates automation, creating fewer and less motivating middle-skill jobs. Conversely, it complements social and innovation tasks, creating more interesting low- and high-skill jobs. This causes labor market polarization, “hollowing out” demand for middle-skill jobs, and increasing wage inequality.