(These are excerpts from my book "Intelligence is not Artificial")
Jobs in the Age of the Robot - Part 2: What Creates Jobs
Unemployment cannot be explained simply by looking at the effects of technology. Technology is one of many factors and, so far, not the main one. There have been periods of rapid technological progress that have actually resulted in very low unemployment (i.e. lots of jobs), most recently in the 1990s when e-commerce was introduced, despite the fact that the digital camera killed the photographer's shop, that Amazon killed the bookstore, that the mobile phone killed the land lines and that Craigslist killed the local newspaper.
The effect of a new technology on employment is not always obvious, and that's why our first reaction is of fear. For example, who would have imagined that the technology of computers (invented for fast computation) would have created millions of new jobs in the sector of telecommunications?
A 2014 report by the Kauffman Foundation showed that between 1988 and 2011 almost all of new jobs were created by businesses less than five years old and "existing firms are net job destroyers, losing 1 million jobs net combined per year. By contrast, in their first year, new firms add an average of 3 million jobs."
New technologies also create jobs in other sectors. It is called the "multiplier effect". The people employed in the new technology need shops, restaurants, doctors, lawyers, schoolteachers, etc. A 2016 report by the Bay Area Council Economic Institute shows that the biggest multiplier effect of our times come from the high-tech industry: for each job created in the high-tech sector, more than four jobs are created in other sectors. A company that hires a software engineer is indirectly creating 4 new jobs in the community. By comparison, traditional manufacturing has a multiplier effect of 1.4 jobs.
Historically, in fact, technology has created jobs while it was destroying old jobs, and the new jobs have typically been better-paying and safer than the old ones. Not many people dream of returning to the old days when agriculture was fully manual and millions of people were working in terrible conditions in the fields. Today a few machines can water, seed, rototill and reap a large field. Those jobs don't exist anymore, but many jobs have been created in manufacturing sectors for designing and building those machines. In the USA of the 19th century, 80% of jobs were in agriculture; today only about 2% are. Yet it is not true that the mechanization of agriculture has caused 78% of people to remain unemployed. Few peasants in the world would like their children to grow up to be peasants instead of mechanical engineers. Ditto for computers that replaced typewriters and typewriters that replaced pens and pens that replaced human memory. Gutenberg's printing press put a few thousand scribes out of business, but it generated a mass production of books, which, besides educating the public and creating an infinite number of new businesses for educated people (e.g. magazines and newspapers), created millions of jobs to print books, market them and sell them. For each scribe that went out of business, thousands of bookstores popped up all over the world. Steam engines certainly hurt the horse and mule business, but created millions of jobs in factories and thousands of new businesses for the goods that could be made in those factories.
All of these forms of automation had side effects that were negative, but one negative side-effect that they did NOT have was to cause unemployment. They created more jobs than they destroyed, and better ones.
In the 1980s i worked in Silicon Valley as a software engineer and back then the general consensus was that software engineering was being automated and simplified at such a pace that soon it would become a low-paid job and mostly exported to low-wage countries like India. Millions of software jobs have in fact been "offsourced" to India, but the number of software developers in the USA has skyrocketed to 1,114,000 with a growth rate of 17% and an average salary of $100,000, which is more than twice the average salary of $ 43,643 (source: US Bureau of Labor Statistics, 2015).
It is true that the largest companies of the 21st century are much smaller than the largest companies of the 20th century. However, the world's 4,000 largest companies spend more than 50% of their revenues on their suppliers and a much smaller percentage on their people (as little as 12% according to some studies). Apple may not be as big as IBM was when it was comparable in power, but Apple is the reason that hundreds of thousands of people have jobs in companies that make the parts that Apple's products use.
I would be much more worried about the "gift economy": the fact that millions of people are so eager to contribute content and services for free on the Internet. For example, the reason that journalists are losing their jobs has little to do with the automation in their departments and a lot to do with the millions of amateur "bloggers" who provide content for free on the Internet.
If we take into account the global effects of automation, we reach different conclusions about the impact that robots (automation in general) will have. In the USA robots are likely to bring back jobs. The whole point of exporting jobs to Asia was to benefit from the lower wages of Asian countries; but a robot that works for free 24 hours a day 7 days a week beats even the exploited workers of communist China. As they become more affordable, these "robots" (automation in general) will displace Asian workers, not Michigan workers. The short-term impact will be to make outsourcing of manufacturing an obsolete concept. The large corporations that shifted thousands of jobs to Asia will bring them back to the USA. In the mid-term this could even have the secondary effect of putting Asian products out of the market and of creating a manufacturing boom in the USA: not only old jobs will come back but a lot of new jobs will be created. In the long term robots might create new kinds of jobs that today we cannot foresee.
Let's take a simple example, a kind of robot that will appear soon at the supermarket of your neighborhood. As you enter the store, you will be welcome by a mobile robot equipped with a basket and a check-out system. The robot will ask you what you are looking for, and escort to the correct aisle of the store. It will let you browse the shelf and pick the brand you prefer. Then it will ask you to drop it in the basket. This will continue for as long as you have items to purchase. In fact, you could even read your shopping list to the robot from the beginning and the robot will calculate the optimal route within the store. For any product that you didn't find the robot will investigate on the spot whether it can be ordered for you and delivered at your home address, or whether there is an affiliated store nearby where you can find it. The robot will also alert you to products that are on sale and politely ask you if you'd like to take a look at them. When you are done, you will simply tell the robot "How much?" The robot will already know the total because it will have scanned each item that you dropped in its basket. The robot will take your payment (whether a credit card or a smartphone app) and print a receipt while escorting you back to your car. A robot like this is perfectly feasible today, except that it would still cost too much to build and operate, a cost not justified in stores that have a relatively low margin of profit on the goods that they sell. Nonetheless, who should panic at the prospect that such robots will someday exist? Which jobs are at risk? There is no human being who performs this task today. When you enter a store, you are on your own. If you have a question, good luck finding any employee who can help you. In many places the check-out operation is already a self-checkout. Hence, not a single job will be lost to these robots. On the other hand, imagine how many jobs will be created. Some company will become the Apple of shopping robots, and hire thousands of people to design, manufacture, sell and maintain these robots. Another company, the Google of robots, will come up with the idea of making the home robot for shopping, a robot that you keep in the house and drives with you to the store, and that knows the organization of each store connected to the cloud according to some Android-like standard. While you are walking into the store with your robot, your robot downloads the configuration of the store and the entire database of products, and then it starts behaving exactly if it were that store's shopping robot. More jobs created here. Another company, the Oracle of robots, will come up with software that informs your robot of the best place for a given shopping list. You won't even know where you are going. The robot will take you to a selection of stores that have what you need at the best prices. More jobs created. Then the Tesla of robots will come up with a way that you can 3D-print your custom shopping robot, as big or small as you want it, as environmental as you want it, as fast or slow as you want it. None of these jobs exist today. And almost no existing job is killed in this simple example.
Not many people in 1946 realized that millions of software engineers would be required by the computer industry in 2013. Robotics will require millions of robotics engineers. These engineers will not be as "smart" as their robots at whatever task for which those robots were designed just like today's software engineers are not as fast as the programs they create. As i type, Silicon Valley is paying astronomical salaries to robotics engineers and China is hiring thousands of people for the Internet of Things.
At the end of 2015 both the McKinsey report "Four fundamentals of workplace automation" and the study by James Bessen of Boston University's School of Law "How Computer Automation Affects Occupations" (November 2015) showed that robots will steal your job but will also create another job, and most likely it will be a better one in terms of income, health and personal satisfaction.
Bessen mentions the case of the ATM, one of the most successful programs to replace humans with machines. The routine jobs of bank tellers were very easy to automate, and, sure enough, the percentage of tellers in the USA fell from 20 per branch in 1988 to 13 in 2004. However, the banks reinvested the money that they saved and, in particular, they opened many more branches, that in turn hired many more tellers. ATMs ended up creating more jobs for tellers, and a huge number of jobs for companies building and maintaining ATMs, jobs that did not exist before.
Unfortunately, you do sell a lot of copies if you write books about the apocalypse, so writers are under pressure to be as negative as possible. Hence, bestsellers such as Martin Ford's "Rise of the Robots - Technology and the Threat of a Jobless Future" (2015) or Jerry Kaplan's "Humans Need Not Apply" (2015), which i thought were misleading, superficial and not helpful at all (to be fair, they both came out before those two reports).
According to the International Federation of Robotics, in 2016 the countries with the highest density of robot population were South Korea (478 robots per 10,000 workers), Japan (315) and Germany (292). These countries also had some of the lowest unemployment rates in the world. For the record, the number for the USA was 164 (the USA has higher, not lower, unemployment than South Korea and Japan), and the number for the euro-countries afflicted by chronic unemployment (such as Greece) was the lowest in the developing world. Data about Italy can be misleading: Italy has a relatively high density of robots but also high unemployment. However, almost all the robots are in the industrialized north, where unemployment is very low, and almost none in the south, where unemployment is one of the highest in the world. The number of robots sold in the United States increased by 43% in 2011 and has continued to increase rapidly, and unemployment has declined every single year since 2011. How in heaven did those distinguished writers conclude from these data that robots cause unemployment?
It is always easy to imagine which jobs will be destroyed and very difficult to imagine the new jobs that technology will create. So we exaggerate the reality of the disappearing jobs and underestimate the reality of the new ones.
In 1992, one year after the invention of the first Internet browser, when newly elected president Bill Clinton assembled a group of experts to discuss the future of the economy, nobody mentioned the Internet (David Leonhardt, "The Depression - If Only Things Were That Good", New York Times, 2011).
The society of robots will create new jobs that today we can't even imagine. Robots will create an even more complex society in which human intelligence will be even more important. The future always surprises us.
And my guess is that robots will become obsolete too at some point, replaced by something else that today doesn't even have a name. Some day robots will be made obsolete by a new human invention. Robots will become obsolete way before humans become obsolete.
Back to the Table of Contents
Purchase "Intelligence is not Artificial"