“The technological progress is deeply unnatural. If it was natural, the industrial revolution would have happened earlier”, says Carl Benedikt Frey in an interview by Robert Siewiorek
Robert Siewiorek: Let us go back 200 years and let us imagine that today is 27 September 1819. What world do we live in?
Carl Benedikt Frey*: We are certainly not in a Sheraton hotel by the beach. Most of the people work in the fields and in agriculture, some work in factories and some work in coal mines. Sometimes those in the coal mines do not see daylight even once a week. Cave-ins and disasters caused by explosions are a part of every-day life. Lung diseases are part of a “working package”. People working in the fields are exposed to all hazards of nature. There are no air-conditioned rooms for hot summer days, there is no roof where one could find shelter from rain. Since there are no machines, only animal power is used on the fields.
What was the meaning of “modern technology” at that time?
At the time, that meaning referred to key technological developments related to railway, which began to slowly spread across Europe. The first railroad was built in Great Britain in 1830. Steam engines were being adopted on a larger scale. That was just a beginning of a real impact of railway on economy in Europe, at least in the western part of Europe. Most people were gradually beginning to see the benefits of changing technology because in other periods of industrialization wages were stagnant or were falling for certain groups on the labor market. So, essentially, after 1840 and after the British economy had taken off, other countries, like France, began to industrialize more rapidly as well.
Who was the creative class at that time?
Bourgeoisie. They were the ones to facilitate the rise of industry. People from that social group achieved what had never been achieved before. The progress that has taken place since then has been enormous. Essentially, due to the rise of merchant industrial class, which made it possible for the progress, with new trade routes and the expansion of modern industry, to spread across the world.
What is a technology trap and how can we avoid it?
The “technology trap” refers to the fact that the technological progress is deeply unnatural. If it was natural, the industrial revolution would have happened earlier. If it was inevitable, then every country would have adopted the same technologies and every country would be rich. What the “technology trap” means is that for most of history people has been a threat to any technology that has threatened their skills and income. And, as a result of that, growth was low and stagnant. The question I pose in my recent book “The Technology Trap: Capital, Labor, and Power in the Age of Automation” is “Could we return to this technology trap in the future in the industrial West?”.
In your paper entitled “The Future of Employment” you and Michael Osborne predict that 47 percent of working Americans may lose their jobs in the coming decades due to progressing automation. Should the Europeans expect the same?
In the paper we tried to estimate the scope of automation. We have found that in the case of about 47 percent of American jobs, the tasks connected with those jobs will not be difficult to automate. The percentage is the same or ever high across Europe, which means that a significant share of jobs done by the Europeans may be exposed to automation as well. However, in Europe the transition to automation on a large scale may be somewhat smoother because the income distribution is more compressed. In consequence, people trading down from middle income jobs that are being automated, like the job of a truck driver, are likely to drop down into service jobs that do not require that much accumulation of human capital. Let us also remember that such jobs are likely to pay more in Europe than they do, in relative terms, in the United States.
Rodney Brooks, a pioneer in robotics, claims that this kind of prognosis is “hysterical”, because the employment depends not only on automation but also on ageing societies. Since the West is ageing rapidly, western countries should invest in automation to fill a labor market gap. Do you think your opinion is hysterical?
Well, Brooks has clearly not read our paper. I think automation is a good thing. Automation is a reason why we are much richer today than we were in 1819, as you mentioned in your question about the past. Thanks to automation we work in air-conditioned offices and not in the fields, factories or coal mines. I think there is no disagreement between us there.
As you mentioned, as populations age and as labor shortages emerge, automation will be essential to maintain our standard of living.
There is a tendency to think that labor shortages will lead to creation of new jobs and that everybody will benefit from that.
And that is just false! It is not even historically accurate. If you look at the first industrial revolution, you will see that wages of a significant share of population were stagnant or even falling for seven decades! If you look at prime age men in the United States with more than a high school degree, you will notice that wages have been falling for four consecutive decades as America is deindustrialized. Let us have a look at prime age men with no more than a high school degree. They are more and more likely to be found outside of the workforce. So, it is just unserious to suggest that this is not a big challenge. Rodney Brooks should read a bit about history and would do best if he started off with our paper.
You claim that we should start worrying about inevitable growing discontent with the side effects of development of digital technologies, because the discontent may cause reluctance to technology itself, leading to deep social and political upheavals. Are we in danger of new Luddism?
I think we are. If we look at one of the research surveys regarding this issue conducted in the United States, we will see that the majority of Americans want stronger restrictions pertaining to installation of machines in companies. When I began to write my book called “The Technology Trap”, robot tax was not part of the debate. Now it is a part of the debate on both sides of the Atlantic. Bill de Blasio is running for president and recently has been drawing a lot of attention to robot tax. The problem is also recognized by Andrew Yang, while Jeremy Corbyn has been talking about it in Great Britain. So I think that there are limits of something that might be regarded a tech clash between men and machines, which does not mean that people will go out to the streets and start smashing machinery. It would be difficult today to smash power grids as it was done some time ago with textile machines. Secondly, people today have the right to protest. They can show their frustration with sticks and stones, and shop on an election day instead of going to a polling station. And that is what they are doing.
There are limits of something that might be regarded a tech clash between men and technology, which of course does not mean that people will go out to the streets and start smashing machinery
Populism is growing both in Europe and in the United States and has to do with industrialization, which is, at birth, a consequence of automation and globalization. It has to do with the old working class or, if you would, the old generation of the working class, who are falling behind and are being politically disenfranchised. While today on the political left there are more people with higher education than in the past and on the political right there are more people with higher income, the populists are very effectively tapping into the anger of the increasingly numerous and politically disenfranchised members of the working class.
In his book “Rise of the Robots” Martin Ford wrote that the middle class would soon become unemployed, that economic mobility would disappear, and that automation would be beneficial mostly for rich technological plutocracy consisting of exclusive elite communities. Unemployed masses will only have basic income and inactivity. Is he right?
No, I think he is wrong. First of all, one of the virtues of democracy is that people can use the democratic process if they are unhappy with what they are offered. And people are already using that process.
Secondly, the worst things that may happened to labor force include slow productivity growth and replacement of labor force with technology. This is the period we live in now. We know that implementation of new technologies, rethinking how the cooperation between people and technologies should look like, and unleashing productivity growth to generate higher wages will take some time. So I think that there is a brighter future ahead that will result from productivity growth. After having made some adjustments and after having acquired new skills, people from the next generation will move to new productive sectors of economy and will benefit from the new situation at a larger scale. Now we are going through a painful period of adjustment. I do not believe in that dystopian long-term scenario painted by Martin Ford in his book.
And what if, as a result of technological progress, intelligent and increasingly efficient machines start to complement not only the work of people but also the work of other machines? At some point, it may turn out that the work we share with the machines will be distributed between the machines themselves, as our work will become completely inefficient if compared to the work performed by the machines.
Yes, what if? First of all, there is some disagreement about the feasibility of general artificial intelligence. There are some expert surveys suggesting that it can happen in twenty years. In my opinion, one of the reasons for that is the fact that if you are an academic and if you want to apply for research funding, and if you say that it is one hundred years away, you will not get any funding.
If you claim that your assumption will prove true in five years from now, you will have to prove it. So twenty years away is just the right time span to make it sound urgent, and you will not have to prove anything.
I do not think that general artificial intelligence would be created as early as in twenty years. We do not even know what path to that scenario would be.
Przeczytaj polską wersję tego tekstu TUTAJ
*Dr Carl Benedikt Frey – economist, historian and writer works at the University of Oxford. He specializes in artificial intelligence economy, history of technology, urban planning and research of digital revolution mechanisms. Being an advisor and consultant, he has cooperated with many international organizations, including the European Commission and the United Nations. His comments are published on BBC, CNN, “The Economist”, “Financial Times”, “Wall Street Journal”, “New York Times” and “Time Magazine”.
The interview was given on 27 September during the 9th Edition of the European Forum for New Ideas in Sopot. Carl Frey was a panelist and one of the guests of the Forum. The sztucznainteligencja.org.pl website was one of the media partners of the event.