by senior futurist Richard Worzel, C.F.A.
“There’s nothing you can do that can’t be done…
Nothing you can make that can’t be made”
– “All You Need Is Love”, Lennon & McCartney
Our own brilliance and ingenuity is putting us out of work.
In some ways, one of the most insidious problems we are facing is the disappearance of meaningful work. Not having meaningful work is soul-destroying for a number of reasons, but two in particular. Not being able to provide for yourself and your family erodes your sense of self-worth and self-confidence. And not having anything meaningful to occupy your time and attention can lead to a loss of purpose and meaning. It’s why newly retired people often die shortly they quit work.
The gradual dwindling of meaningful work has been happening for decades, but is now accelerating. Mostly, it’s happened for two reasons: foreign competition, and domestic automation.
Foreign competition – the migration of jobs to developing countries like China, India, and others, where the workers do effectively the same work for a fraction of the cost – has been much discussed, and continues to receive a lot of attention (although much of this is disguised as discussions about trade rather than work). I’ve discussed this before, and will again. Now, though, I’d like to focus on domestic automation.
The Rise of the Robots
It’s clear to me that eventually robots and computer intelligences will be able to do almost anything that humans can do, and do it better, faster, and cheaper. The science fiction writers of the first half of the 20th Century saw this as leading to a utopia, where everyone lived as they wished, free of the drudgery and toil of back-breaking labor. Indeed, I still get asked about this future quite regularly. I call it the “George Jetson future.”
George Jetson, of the animated TV show “The Jetsons”, was employed by Spacely Sprockets to fly into his office sometime around 11 a.m., and sit with his feet up on his desk until noon. Then, when the noon lunch whistle went, he would push a button in the middle of his desk, and go home for the day. That’s the old science fiction utopia.
No one thought to ask: If all George Jetson does is push a button when the noon whistle sounds, why do we need to employ George Jetson? Once you ask that question, the answer becomes obvious: you don’t. And that answer leads to the world we are living in today, rather than the George Jetson future.
Machine learning, coupled with the continuing acceleration of computing power and the increasing sophistication and subtlety of computer techniques, is rapidly creating machines that can learn new tasks, and perfect them, slowly at first, but at a rapidly accelerating pace. Moreover, this is happening in many fields, with each one perhaps using a different set of techniques. Anyone who cares to do even casual research can verify that this is happening, and in a growing number of areas of what used to be human work.
That doesn’t mean that we are creating a race of humanoid robots that look, talk, think, and act like humans, only better. Robots will come in many shapes, and with differing abilities and specialties. And computer intelligences will be focused on specific tasks, rather than being able to do a little of everything, as humans do. But collectively, the result is a range of robots and computers that are smarter, faster, and can do a growing number of things more cheaply than humans.
The Neo-Luddites vs. the Technologists
This leads us to a very old debate, dating back at least to 1811, when textile workers rebelled against the mechanical automation of weaving and spinning machines in often violent protests. These people were called Luddites because they were believed to be followers of one Ned Ludd, who was reputed to have smashed such equipment in protest. The term Luddite is still widely used, and generally applied to people who are stuck in the past, and rail against beneficial automation. Their argument, then and now, is that machines are eliminating human jobs, and hence the advance of such machinery should be stopped.
On the opposite side of the debate are people who have no such catchy label, but which, for convenience, I will call the Technologists. Their argument, which has been right for the last two hundred years, is that as automation eliminates bad, old jobs that rely on repetitive actions and backbreaking labor. As a result, automation increases productivity, which decreases the cost of goods and services produced. In turn, this increases the standard of living by leaving people more money to buy more goods and services, and, along the way, creates new jobs by increasing demand. And as a result of this rising demand, new jobs are created that are better paying, offer better working conditions, and greater hope for advancement.
In essence, new, better jobs are built on the rubble of bad, old jobs. And for the past two hundred years this has been proven to be correct.
But it may not be as true today, and it certainly won’t be true in the future. Automation still increases productivity, which still lowers the prices of goods and services. That still increases the standard of living – but only for those who are gainfully employed, and have the money to buy things, and that’s where the Technologists’ argument is finally breaking down.
Automation Escapes the Factory Floor
As computers have become more powerful, and more sophisticated, they are displacing workers in jobs that not only involve muscle power, but now, increasingly, white collar workers and professionals that are involved in repetitive tasks involving thinking and judgment.
IBM’s Watson computer system is perhaps the best known, and possibly the most advanced, of such systems. Watson became famous in 2011 for defeating the two human champions of Jeopardy!, the television game show that makes use of the eccentricities of human language, in a three-day match. IBM’s purpose in creating Watson was to produce a machine-learning system that could understand human language, absorb enormous amounts of information, and come up with answers, based on weighted probabilities, to certain kinds of questions. The first area, for instance where Watson was applied was in medical diagnostics, starting with certain kinds of cancers that are difficult to diagnose.
Doctors are not being replaced by Watson, but their work is being supported, and, if you want to be cynical about it, double checked by Watson’s diagnostic abilities, especially in particularly complex diagnoses.
BakerHostetler, one of America’s largest legal firms with over 900 lawyers, recently engaged the services of ROSS, a computer application which was developed by a company called ROSS Intelligence. ROSS makes use of Watson to perform legal research for law firms like BakerHostetler. ROSS provides extensively researched legal briefs that have been described as impressive, and gets progressively better as the lawyers that use its output accept or criticize with the material it presents them. As a result, BakerHostetler won’t need to hire anywhere near as many new lawyers or articling students. And ROSS will continue to get better and better as time goes on.
Accountants and tax preparers are finding that the routine aspects of their work is increasingly being done by computer systems, which means we need fewer accountants and tax preparers.
Customer service representatives (“CSRs”) that answer phone systems are increasingly being replaced by computer systems of growing sophistication that can answer most routine questions or problems, and refer the rest to human CSRs, thus reducing the number of humans required.
Nor are office workers and professionals the only ones being affected. Restaurants are looking at computers and robots to replace minimum wage workers, both for taking orders (as is starting to happen at McDonalds), as well as for cooking food (as with the robot burger-maker made by Momentum Machines).
But What About the New, Better Jobs?
But what about the new, better jobs that Technologists say are created, that offer higher pay and more opportunity? Well, some of that is happening, but it’s no longer as certain as it was in the past.
New jobs are being created, but they tend to come in two varieties. High-level jobs are being created that pay well, but they tend to require exceptional levels of specialization, plus high levels of education, usually including a college degree, and often requiring post-graduate qualifications as well. That immediately excludes the large majority of the population without qualifications beyond high school, as well as those with the wrong kinds of college degrees, or without the particular specialization required.
Worse, the shelf-life of such jobs is growing shorter as the pace of change accelerates. Hence, today’s high-level, well-paying job may disappear within less than five years, and possibly less than two.
As a result, while new jobs are being created, they offer less job security than in the past. And once one of these jobs disappears, the person thrown out of work often does not have skills that match up with the even newer jobs created. This can make them as unemployable as a displaced auto worker. And if you look at the U.S. Bureau of Labor Statistics chart below, you can see the net effect: fewer and fewer people of working age are employed, reversing a trend established after World War II. (Note that the recent upward blip is a cyclical response to the improving U.S. economy, and doesn’t represent a reversal of this long-term secular trend.)
The second kind of new jobs that are being created are low-paying service jobs, typically called McJobs. Even people with high-level skills can wind up taking these kinds of jobs simply because they no longer have skills that match up with today’s jobs. And, as mentioned earlier, these McJobs are also being whittled away.
What Happens Next?
What happens next is a question that has a number of different answers.
First and most directly, automation is going to continue to eat its way up the employment food chain, replacing more and more workers, and in increasingly complex and sophisticated kinds of work. Indeed, I would suggest that any kind of routine work will be automated sooner or later. Hence, if your job involves doing the same thing, or similar things, over and over again, you are at risk of being replaced by a computer intelligence or robot, regardless of whether you are on the factory floor, do white collar work in an office, flip burgers, write music or novels, or are a professional doing work that requires extensive post-graduate training. If your work is repetitive, you risk being replaced.
But if almost all routine work is replaced, what does that leave?
What’s left is mostly non-routine work, or creative, innovative work. This is harder to replace because each day is unique, and tomorrow’s work often doesn’t exist until you define it. This is difficult because you are constantly inventing the future, and figuring out new reasons why your clients (or your boss) will want to keep paying you. Most, but not all, of my work falls into this category, for instance.
And I would argue that all work done by humans in the future will either be creative in some way, or you will be working with, most likely serving, people who want to be served by humans, not machines. Hence, high-end restaurants are likely to retain human servers and human chefs because the patrons don’t want to pay fancy prices for a romantic dining experience, for example, only to be served by a robot.
Likewise, computers are likely to do a lot of the heavy lifting when it comes to medical diagnostics, which involves sifting through large amounts of often contradictory data, but those aspects of medical care that require a human touch – a bedside manner, or even compassion, if you will – will continue to be done by humans.
Hence, it may be that nurses are more likely to have job security than many doctors. But doctors will still be critical to health care because of their awareness of the human elements of patient care, and people want to know that a competent human is in charge, much as airplane passengers want to know that there’s a human pilot in the cockpit. The maitre d’, or concierge of an upscale restaurant, hotel, or nightclub might also fall into this category.
Artists and crafts people who make unique objects (paintings, sculpture, furniture, woodwork, and so on) involving great skill will largely be immune, especially in the case of objects that have a certain snob appeal. For instance, you are unlikely to be willing to pay as much for a copy of a great painting, made by a very skillful robot, as you are for the original, painted by a great painter.
Would you be willing to buy a novel that’s a great read for an airplane if it was written by a computer? I suspect the answer is yes, but I also doubt if computers will be able to write great literature, or come up with truly original ideas.
So, job security will come in three primary forms: those that involve doing new things most of the time (which has its own insecurities); those involving products that are themselves the result of great creativity, especially where there is a cachet in owning something that is made by a specific human, typically an artist or craftsperson; and those where a human touch or human service is particularly prized or important. In other words, our salvation in the face of automation will be the things that make us uniquely human.
But stop a moment: do our current education and employment systems recognize the need for such humanity and creativity? Are we helping the workers of today and tomorrow improve their chances of employment? Not from what I can see. It seems to me that we are still running schools based on 20th Century, mass-production models of education and employment, and pushing people to seek jobs that may have limited shelf lives.
So everyone’s job is at risk if it’s based on repetition. Which brings us to the other major aspect of what happens next.
In some ways, I’m surprised there hasn’t already been more of a backlash against the machine. Yet, there are explanations why it hasn’t happened so far.
First, this is a gradual change that has been happening slowly up until now. But now the pace of change is accelerating as technology accelerates.
Next, the job losses to automation are buried in other effects, such as the high unemployment that occurred during the Great Recession of 2009-10, and the gradual recovery as the economy has improved since then. These effects, which are well-known going into, and coming out of, recessions tend to get talked about because they are familiar. By comparison, displacement of humans by machines at the speeds I’m beginning to see is a new thing, and people have to notice it before it gets much attention. That, too, is changing as the effects become more evident, especially as robots in particular are sexy, and make good media copy.
Then again, the job losses to foreign competition have garnered major headlines for so long that there’s a natural tendency to ascribe similar changes to the same cause. It’s just easier to do that than to dig up a new reason and verify it.
As well, employers don’t make a big deal about replacing humans with automation. They know it makes them look bad in the media, even though it may make their financial results look good. Plus the political effect of automation is to improve the position of the so-called 1%, who own the machines, at the expense of the 99%. The rich and powerful tend to have great economic, marketing, and political clout.
And one final reason why this hasn’t garnered more attention yet: people who are out of work tend to have other things to worry about than commenting on socio-economic changes. They’re worried about being able to pay the bills, and about finding another job. And they and their views tend to be discounted as they are relatively powerless, and may even be deemed unimportant in our society.
But eventually, people will take notice of the encroachment of the machines, and how it is affecting them, their families, or their friends and neighbors in a significant way. It’s at that point that a backlash will start.
Indeed, we may already be seeing a bit of a backlash. In supermarkets, shoppers often line up to be checked out by a human cashier than go to the (admittedly balky and difficult) self-check-out machines. Some telecom companies are going back to humans to field service calls over the phone, and making marketing hay out of it.
Whether this is just because the technology isn’t good enough yet or not is unclear. But there are times when people don’t want to deal with a machine. And I believe that we will start to see slogans along the lines of “We don’t buy from machines!” I expect that there will be a movement against machines comparable to the movement against GMO foods.
The effects of replacing humans with machines are great, and growing. Robots and smart computers can often replace a human for a one-time cost that is equal to, or less than, one year’s wages for a human. As that cost continues to come down, and as the capabilities of robots and computer intelligences continues to grow at accelerating rates, the case for robots will become increasingly compelling.
The Cost of Being Cheap
In the short run, and for an individual employer, the case for the machine may be compelling because it makes so much economic sense, and does so much for the bottom line. Yet, what’s true on the micro-scale isn’t true on the macro-scale, because collectively employees are consumers, and the more employees our economy eliminates, the fewer consumers will be able to buy the goods and services produced. So what works for an individual company does not work for the economy as a whole.
Left unchanged, this could lead to revolution, humans against the machines and their owners. Certainly concentrating wealth in the hands of a few, and economically disenfranchising the many, has historically lead to social unrest, and leaves us ripe for loud-mouthed leaders who rail against the establishment, and offer simple-sounding solutions to very complex problems, yet have no real insight about what to do. That way lies despotism and bloodshed.
Is there another alternative? Perhaps.
Productivity comes either from using fewer employees to produce the same number of goods and services, or from producing more goods and services from the same number of employees. So far, most producers have chosen the first alternative. But what would happen if we chose the second?
Well, more people would be employed, but the price of goods would fall even more rapidly as supply expanded, increasing the standard of living while maintaining higher levels of employment. A larger number of people would be better off, as would the economy as a whole. It could (and probably would) be argued that business owners might not make as much profit as they would with fewer employees as other costs (real estate, cost of materials, and so on) wouldn’t decline, but they would be more likely to have a greater number of affluent customers. Remember that Henry Ford gave his workers an unprecedented raise in wages so they could buy his cars, even though on paper, it raised his costs, and didn’t seem to make sense.
Is this a feasible solution? Again, it might be, but only if companies and their employees look for ways of re-deploying people who are made redundant, rather than just letting them go. Hence, instead of getting a pink slip, you might get a blue slip which says, “We’re replacing your job. For the next three months, we’d like you to research ways that you can continue to be productive with the company. Come up with a way to invent a new job for yourself, and then we’ll see if we can make it work.”
This isn’t easy, and it wouldn’t work all the time, or for all workers. But it is, to my mind, a more sustainable solution to a problem most people aren’t yet aware we have.
So: pink pill, or blue pill. Which future should we choose?
© Copyright, IF Research, May 2016.
N.B. For another viewpoint and more in-depth review of this subject, read Martin Ford’s new book, Rise of the Robots: Technology and the Threat of a Jobless Future