Will Robots Do Our Work?

robot, work, unemployment, WALL-E

I contributed this article to the October 2017 edition of THink Magazine. The issue, “Challenges of Our Time,” looked at big picture challenges facing humanity. My piece focused on long-running trends in the mix and availability of work, and the technological waves that drove the shifts in those trends in thinking about whether “robots” might be looking to steal our gigs, as well as what that might lead to.

“The difficulty lies not so much in developing new ideas as in escaping from old ones.” – John Maynard Keynes

Few things are universally revered. Family, friendship, music, sharing meals, and a strong work ethic are a few that come to mind. The last one seems the odd one out in the list, but it has long been in service to the rest. If you wanted to spend quality time with family and friends, in comfortable homes or public spaces, while being entertained or enjoying a nice meal, you had to work for it.

For millennia, everything was done by human hand and so a good work ethic was an important value to ingrain into society. The field was not going to sow itself, nor would the house be self-built. Each had to pull their own weight to ensure their basic needs were met.

Since the first stone tools were used, technology has been augmenting work. It typically creeps in at a pace that is not broadly disruptive, but it occasionally lurches forward in periods we call “industrial revolutions”. Three such phases (mechanisation, mass production, and automation) have brought humanity great advances. But with each revolution, the mix of jobs shifted as technology took over existing work and helped foster new jobs. In the West, it started with work in the primary sector (agriculture, mining, and fishing) being transformed by relatively simple, but powerful machines.

technology, unemployment, robots, ai
Breakout of US labour force by sector.

The secondary sector (manufacturing) followed a similar pattern, so that most of the workers in those economies now work in the tertiary sector (service). Some researchers have further recognised a quaternary sector which consists of research and development. The distinction may seem like splitting hairs, but it is interesting to note that at least for the United Kingdom (UK), this fourth sector effectively accounted for all relative job growth towards the end of the 20th century.[2]

The secondary sector (manufacturing) followed a similar pattern, so that most of the workers in those economies now work in the tertiary sector (service). Some researchers have further recognised a quaternary sector which consists of research and development. The distinction may seem like splitting hairs, but it is interesting to note that at least for the United Kingdom (UK), this fourth sector effectively accounted for all relative job growth towards the end of the 20th century.[2]

technology, unemployment, robots, ai
% of UK employment by sector 1800-2000

We may now be entering a fourth industrial revolution wherein robotics and artificial intelligence threaten to greatly alter our economies. How they will impact society largely remains to be seen, but there is growing concern of mass job loss as highly adaptable machines and software could creep into areas that have long seemed safe from technologically driven displacement. The quaternary sector may explode, but can it keep pace with the jobs it destroys?

Human labour (both manual and cognitive) is hardly an endangered species. But if a significant portion of it quickly evaporates thanks to technological advances—in an era when things like incomes, healthcare, and retirement savings are often directly tied to jobs—how will we react?

Take the United States (US) trucking and taxi sectors for example. Driverless vehicles threaten to erode into these millions of jobs, while also delivering a significant hit to the supporting industries and the communities that truckers and taxi drivers visit along their routes. A doomsday scenario in which these jobs evaporate overnight seems farfetched, but what might it do to a sizeable city if the local fleet of cabs or the “ridesharing” vehicles were all hurriedly replaced? Assuming the costs continue to come down, and the capabilities continue to increase, is it realistic to think—barring unforeseen technical issues or possibly a high-profile accident—that the technology is not primed for proliferation, and human labour headed for the curb?

Robots have been putting pressure on employment and wages in the US since at least 1990.[3] The effects have been limited so far, but these machines, and their counterparts in artificial intelligence, appear to be just getting warmed up. The question we need to ask ourselves is, “How will we like to react to it if they really take flight?” If we think through the possibility, we will be better prepared to react should it occur.

Towards uncharted territory?

Today’s productive benefits from technology, and its encroachment on employment, have long been expected. John Maynard Keynes, the eminent economist of the early 20th century, warned us of this nearly a century ago. He noted the change from an average work week had dropped from around 60 hours in the mid-19th century to around 47 in 1930. So, at the time, it probably seemed obvious that a short work week was well on its way.[4]

technology, unemployment, robots, ai
Working hours for the UK and US (1870 to 2000)[5]

The data for the US between the 1930s and 1940s is skewed (presumably thanks to the Great Depression and WWII), but from 1950 to 1970, the downward trend continued, as it dropped from about 42 hours per week down to around 38. Had that trend continued, we would be nearing a 30-hour work week. But the figure bottomed out in the 1970s, and it has hovered around 40 ever since. Meanwhile, productivity rose unabated.[6]

If these trends continue, and technological progress accelerates, we may be headed toward some interesting challenges. Longer work weeks have long been favoured by employers for their ability to reduce the associated costs of benefits. Replacing human labour with technology can afford similar savings, while also offering a host of other benefits (at least from the view of the employer). Making such changes are rational choices for firms that are expected to maximise returns. But aggregated up to a macro perspective, this can become problematic. If technology progressively erodes work and the collective response is to employ fewer workers (and new jobs are not created to replace the ones that have been lost), the burden that is created will fall disproportionally on the unemployed individuals and, to a lesser extent, society at large.

Technological advances might provide significant boosts to economies, but they may also decrease social mobility—a measure of the movement of individuals between a society’s social strata—which could lead to a smaller economic pie. The UK lags just behind the average measure of social mobility for Western Europe. By bringing their society in line with the region’s average, it is believed that it would bring an associated increase in annual GDP of around 2%—equal to an additional £590 per person.

It is no surprise to see firms like Facebook and Google put massive investments into the development of new technology. But when an auto-manufacturer like Toyota jumps into the deep-end with US$100 million for technological investment, we are either seeing a phase of irrational exuberance on the investment side, or those making the investment decisions see that we are on the leading edge of something big.

If technology does get into a new phase of leaps and bounds, it is possible that there will be pushback.[7] If others dislike interacting with newly automated systems as much as I disdain telephone voice response systems, maybe technology will be held to certain quarters, and maybe we will demand human interaction. But if we take recent decades as a guide, whether customers are happy about such changes, does not appear to be the deciding factor. If so, we are probably faced with having significantly less work to go around.

A simple thought experiment might help us prepare for that possibility. Imagine what might happen if the displacement of work accelerated and had significant impact over a reasonably short timeframe. The work would still get done, and possibly significantly more would be accomplished overall. But let’s assume the net effect of the change was that the city you live in lost 10% of its compensated work. Don’t worry about which jobs would be lost. Don’t worry about individual pay levels. Just think of the big picture. What sort of stress might that place on support systems? What might it do to the social fabric? What if the loss of work accelerated from there? Keynes had an answer ready for those circumstances. He felt that we should “endeavour to spread the bread thin on the butter—to make what work there is still to be done to be as widely shared as possible.”[8]

When he gazed into his crystal ball, Keynes saw a future in which he thought we would be working around 15 hours a week. His only worry with that was that we might not know what to do with ourselves—the perils of idle hands. One recent study lends some support to this fear as it found that we have gained about 4 hours per week of leisure time thanks to the automation of some of our household chores. On average, we have largely filled those hours watching television. But unless such viewing is somehow destructive, maybe it’s not such a bad thing. If we could go back in time and ask people whether they would rather manually wash dishes and clothes, or zone out watching TV, I’m guessing the choice would be an easy one for most.

Further, the television viewers were reclaiming non-work hours for leisure. But if we started making significant reductions to today’s work week, we would be looking at a different opportunity, one in which we might encourage people to use some of the time that has been freed up for the collective good. Rather than reducing the number of people working, we could just reduce the number of hours worked, while maintaining our income levels. All would do their part and all could take care of their needs. Because the question is not whether technology will continually take on more of our work—that seems assured—but whether we’ll ask technology to lift us all up, or we’ll allow it to drive us apart.

Looking back through history, the length of a work week is a relative notion, but the 40-hour standard has been around long enough to have been normalised for those who are alive today. For us, changes will not be viewed as the continuation of a long-running trend, but as a break from the norm. This opens the door for treating these hours differently than prior gains. And since we have successfully continued inculcating the need for industriousness long after it was necessary for a reasonable measure of subsistence, could we not redirect those efforts towards socially beneficial aims?

If we ever made Keynes’ vision a reality, I would suggest we encourage community-building with mottos like, “Fifteen for my livelihood. Fifteen for the social good.” Maybe that wouldn’t work for everyone, but a future in which we all have relatively comfortable lives, significantly more free time, and large numbers of us coming together to build community sounds great to me. If we can agree on that, I say we let the robots have all the work they want.

References:

[1] Johnston, L. D. (2012, February 22). Understanding the decline in manufacturing. MinnPost. Retrieved from  

[2] BBC. (2014). Characteristics of industry. Bitesize, BBC website. Retrieved from  http://www.bbc.co.uk/schools/gcsebitesize/geography/economic_change/characteristics_industry_rev3.shtml

[3] Acemoglu, D., & Restrepo, P. (2017). Robots and jobs: Evidence from US labor markets. Cambridge, MA: National Bureau of Economic Research. Retrieved from http://www.nber.org/papers/w23285

[4] Keynes, J. M. (1963). Essays in persuasion. New York, NY: W. W. Norton. See pp. 358–373; Blodgett, H. (2014, December 27). Over the past 150 years, there has been a profound shift in what we do with our time. Business Insider. Retrieved from http://www.businessinsider.com/how-humans-spend-their-time-2014-12

[5] Roser, M. (2017). Working hours. Our World in Data. Retrieved from https://ourworldindata.org/working-hours

[6] Blodgett, H. (2014, December 27). Over the past 150 years, there has been a profound shift in what we do with our time. Business Insider. Retrieved from http://www.businessinsider.com/how-humans-spend-their-time-2014-12; Reich, R. (2011, September 3). The limping middle class. New York Times. Retrieved from http://www.nytimes.com/2011/09/04/opinion/sunday/jobs-will-follow-a-strengthening-of-the-middle-class.html; New York Times. (2011, September 4). Infographic. Retrieved from  http://www.nytimes.com/imagepages/2011/09/04/opinion/04reich-graphic.html

[7] Shead, S. (2017, March 30). Google and Mark Zuckerberg’s investment fund are backing a $150 million AI institute in Toronto. Business Insider. Retrieved from http://www.businessinsider.com/google-is-backing-a-150-million-artificial-intelligence-institute-in-toronto-2017-3; Terdiman, D. (2017, July 11). Toyota gears up against Google, Ford, and Tesla with $100 million AI venture fund. Fast Company. Retrieved from https://news.fastcompany.com/toyota-gears-up-against-google-ford-and-tesla-with-100-million-ai-venture-fund-4043106

[8] Keynes, J. M. (1963). Essays in persuasion. New York, NY: W. W. Norton. See pp. 358–373.


Also published on Medium.