Technological advancement has always been a transforming force, generally thought of as positive, throughout the history of the modern era. Over time, technological advancement has replaced certain jobs and complemented others, and ultimately created more flexible and rewarding work.  Automation is a form of technological advancement that has witnessed massive growth in recent decades.  Robots in particular are continually pushing automation to areas previously unanticipated.  The once “dumb” linear-tasked robots are now giving way to new “smart” robots that can leverage processing power from “the cloud,” untethered by one particular hard drive, and even “think” for themselves via machine learning.  The new smart robots are popping up in our homes, on our roads, teaching in schools, and even serving as coworkers on the job.

Recent developments in robotics have been pushing the envelope faster and further than was expected even a decade ago, making what was once thought of as science fiction a reality.  Whether it is vacuuming our carpets or assisting in precise surgeries, robots are helping with and performing a multitude of tasks.  The robotics market has been growing at a rapid pace, and purchases of both industrial and service robots are reaching record levels.  Venture capital invested in robotics doubled last year, with much of it flowing to smaller firms with high growth potential. 

Automation Contributes to Current Job Polarization Worries

In the aftermath of the last recession, technological improvements in automation and robotics are frequently associated with negative effects today, such as replacing middle-skill workers, a phenomenon known as “job polarization.” With fears of technology replacing humans renewed, many begin to question whether this time technology will replace human labor for good. When referring to the backlash against automation, the most oft-cited story is that of highly skilled English textile workers, called Luddites, smashing looms in the late 1800s as a form of protest against their displacement by machinery.^[1] Economist John Maynard Keynes popularized the term, “technological unemployment” in describing the loss of jobs due to technological change, though the phenomenon itself has been referred to even in Aristotle’s time.[2]

Robotics and automation are expected to exacerbate the current job polarization trend in the United States and other developed countries.  In 2013, University of Oxford researchers Carl Benedict Frey and Michael A. Osborne examined what jobs in the American labor market are most susceptible to automation over the next few decades.  After examining over 700 different occupations, they concluded that 47 percent of American jobs were at risk of automation.^[3]

Even though recent pessimism over the potential negatives of automation persists, many have noted that widespread technological unemployment may be overhyped.  Automation still struggles with many creative and dexterous activities.  When programmers tried to create a program to emulate legendary jazz pianist Bill Evans, it was a resounding failure.^[4]  Further, as the Financial Times noted, robots have difficulties with locating and grabbing wires and take about twenty minutes to fold a towel.^[5]  An additional advantage possessed by humans is our ability to “reprogram” ourselves in a fraction of a second to respond to demands for customization.

 

Automation Continues to Replace, Enhance, and Create Jobs

Many “middle-skill” jobs have been disappearing as routine tasks and processes, regardless of whether they require “brains” or “brawn,” are streamlined.  Some of the occupations are likely to continue to be automated, such as telemarketing and sewing by hand.  However, others like fashion models and restaurant hosts or hostesses, are not likely to be replaced.  Current automation is limited especially in areas that require creativity, dexterity, and “thinking on one’s feet,” as well as the social aspect only found in human nature.  These realms are still allowing humans to compete, and even work alongside, their robotic counterparts. 

A refinement of the skill-biased technological change explanation for shifts in the labor market, described by MIT economists David Autor and Daron Acemoglu as the polarization hypothesis, emphasizes the notion that technological improvements tend to replace workers whose tasks are largely routine. As noted by the Federal Reserve Bank of San Francisco, “Routine jobs are those that follow relatively set rules and consist largely of repeated actions, for which computer technologies are well-suited and can largely replace human labor. By contrast, non-routine jobs require flexibility and often social skills. The second dimension, cognitive versus manual, reflects the mental versus physical focus of the required tasks.”[6] As shown in the figure, routine jobs have been declining since 2007, and have seen little-to-no growth since 2000.

In fact, a McKinsey report notes that instead of widespread “occupational” automation, certain activities within occupations are the most likely to be automated.  For example, most airlines now have computerized kiosks that allow customers to check-in and receive their boarding passes.  However, these kiosks have yet to eliminate the need to have an agent at the desk to perform other more complex tasks that the kiosks cannot handle.  Although the computer enables airlines to save on labor and reduce customer wait times, it has not completely replaced the agent’s job.^[7]

While automation has replaced and enhanced a number of jobs, it has also created a host of new jobs in fields that not even Aristotle nor the Luddites could have imagined, particularly in programming, bioengineering, nanotechnology, medicine, energy, education, and professional services. These high-skill jobs are not only complemented by automation, but would not have existed without automation. They represent the cutting-edge frontier of innovation and come with high compensation for those who acquire the skills to perform them.[8]

Overcoming Job Polarization Trends

FiveThirtyEight’s Andrew Flowers took the aforementioned Oxford University analysis one step further by asking what jobs are most at risk from automation and are held by a majority of workers earning close to the minimum wage.  For those jobs for which at least 25 percent of workers earn $10.10 per hour or less, 82 percent were at risk of automation.  Among the top three minimum-wage occupations by number of workers—retail salespeople, cashiers, and waiters and waitresses—the median probability of computerization is 94 percent.  Already, the rise of “self-checkout” lanes marches on in retail stores nationwide.  The nature and greater potential of automation for these types of jobs led Flowers to conclude, “...it is clear that low wage workers are more vulnerable to computerization.”^[9]

However, other experts have applauded the replacement of these low-skill jobs with robots because it allows workers to move to other, more lucrative jobs where there is a high demand, such as health care or education.^[10]  The demand for these jobs is increasing and the wages are decent, but unfortunately these jobs require education and training that workers may not have.  Furthermore, many come with expensive licensing requirements that could be out of reach for workers.^[11] Rather than hinder innovative workers and the machines that can complement them, government policy can encourage thriving employer-employee relationships through smart regulatory reform that accomplishes two goals:

1)     Reducing barriers that artificially increase  the cost of hiring workers, and

2)     Reforming burdensome regulations that divert business resources to costly compliance.[12]

As noted in Flowers’ analysis, the increasing affordability of flexible robots like “Baxter,” a robot that learns by watching a human perform a task, will begin to put pressure on the lowest-paid workers.  Annual income for a full-time worker earning, for example, $10.10 per hour amounts to just over $21,000, excluding any potential fringe benefits offered or compliance costs and additional taxes associated with payroll.^[13]  A base model of Baxter with a one-year warranty and mobile pedestal is quoted at $34,700.^[14]  Although Baxter is still more expensive than the employee, the costs of its work do not vary much by hour, and it is not subject to labor regulations or overtime requirements.  Further, if Baxter lasts more than one year, the employer does not have to pay any extra money besides electricity and potential maintenance costs.  Scenarios like those above have led certain scholars to label minimum wage increases with a pejorative title, the “Robot Employment Act.”^[15] Instead of creating an artificial, government-imposed disadvantage to automated competitors, policymakers should employ solutions to make it easier for people to obtain and offer work.

Furthermore, in an era characterized by lower workforce participation and slower business dynamism, government policy solutions should focus on removing barriers to employment and entrepreneurship. As Andrew McAfee noted in the JEC hearing, entitled “The Transformative Impact of Robots and Automation,” entrepreneurs are facing an “increasingly dense thicket of things that an employer or worker has to confront before they can start something up. And navigating your way through that becomes increasingly difficult, and it looks like more and more people are saying, ‘I’m just not going to bother with it’.” McAfee lauded the Silicon Valley notion of “permissionless innovation,” which argues for greater ability to innovate without submitting to a bureaucracy or permitting process before proceeding.[16]

Skill Adaptability Hangs in the Balance

While technology has been advancing, America’s labor market has shifted from a relatively young, low-skill, production-intensive juggernaut to an older, more nuanced, services-oriented and high-skill dynamism.  Similarly, our expectations shift going forward.[17]  Workers can take advantage of their exclusively human ability to adapt and reprogram their skills to improve their desirability in the labor market today. Accessing training in the form of massive open online courses (MOOCs) has never been easier or more affordable for learners the world over, thanks in large part to the fruits of decades of technological advancement.  This is in contrast to the usual government retraining programs offered up as a means to reduce “technological unemployment.”  The Government Accountability Office (GAO) has found questionable effectiveness of government retraining programs and considerable overlap.^[18]  On the other hand, MOOCs and other forms of learning through apprenticeships or on-the-job training offer more efficient and flexible alternatives than the existing web of government training programs. 

Beyond “retooling” current worker skills, robotics is challenging Americans to rethink what their education system is teaching future workers. It is problematic that the U.S. education system has developed virtually only one path to success: obtaining a college degree. Not all degrees are equal.  In particular, Science, Technology, Engineering, and Math (STEM) degrees will command higher wages as demand increases to create more machines.  While non-STEM degrees can also have much value to society, we should not assume that any type of college degree leads to success, nor that STEM careers require a four-year college education.  The focus on college has set up many students for failure to complete their degrees, as many may not have belonged on a college track in the first place. Many could have been learning valuable trade skills, which automation has served to enhance and amplify.  For example, high school students in Germany often start apprenticeship programs in high school and graduate better equipped to begin a career than many American high-school graduates.^[19]  Other routes can lead to success, many of which do not even include a bachelor’s degree.[20] Given sufficient and relevant training, many of these education routes can still lead to higher-paying STEM occupations, as projected by the Georgetown University Center on Education and the Workforce, shown in the figure below.[21] An example the U.S. Census Bureau notes is engineering technicians, who often have a vocational degree.[22] Even more workers can be expected to work in STEM-related occupations over the next decade, particularly in the health care sector.

As economist Alex Tabarrok argues in The Chronicle of Higher Education, while there appears to be a need for a greater focus of funding STEM education—which has the potential to confer greater benefits to society through technological innovations—a more pressing need is to focus on students that have fallen behind, including millions of college and high school dropouts.  Tabarrok points out that the “obsessive focus” on attaining a college degree has not served taxpayers or students well.  Given that the United States has the highest college dropout rate in the developed world, it is perhaps problematic that the U.S. education system has developed only one path to knowledge, when there are “many roads to education.”^[23] As Tabarrok concludes, “We need to provide opportunities for all types of learners, not just classroom learners.  Going to college is neither necessary nor sufficient to be well educated.”^[24]

The demand for skilled work is increasing, and as the reality of automation sets in, it will continue to signal what types of skills should be taught in our schools and colleges. A quality education fit for the 21^st century involves providing students with fundamental skills—such as arithmetic, reading comprehension, writing and analytic thinking—as well as strengthening the connection between students and employers. Doing so would empower workers with the skills they need to fill the jobs that employers offer today and make them better equipped and more resilient to the role automation plays in labor market dynamism.