Angela Merkel gestures towards an interactive robot

Economics

Rage Against the Machine?

Dec 7, 2017 Laura Tyson and Susan Lund

Almost every aspect of our economies will be transformed by automation in the coming years. But history and economic theory suggest that fears about technological unemployment, a term coined by John Maynard Keynes nearly a century ago, are misplaced.

BERKELEY – Intelligent machines are transforming the way we produce, work, learn, and live throughout the world. Almost every aspect of our economies will be radically altered.

Major logistics companies and individual drivers are using new technologies to optimize their route planning. Companies like BMW and Tesla have already released self-driving features in their automobiles, which are produced with the help of sophisticated robots. The Associated Press is using artificial intelligence to help write news stories. 3D printers are being used to produce replacement parts – for both machines and humans. AT&T, in collaboration with Udacity, is offering online “nanodegrees” in data analytics. Drones are delivering health supplies to remote locations in poor countries.

These marvelous new technologies promise higher productivity, greater efficiency, and more safety, flexibility, and convenience. But they are also stoking fears about their effects on jobs, skills, and wages. Feeding these fears is a recent study by the University of Oxford’s Carl Frey and Michael Osborne, and another by the McKinsey Global Institute (MGI), which find that large shares of employment in both developing and developed countries could technically be automated. History and economic theory, however, suggest that anxieties about technological unemployment, a term coined by John Maynard Keynes nearly a century ago, are misplaced.

To continue reading, register now.

Subscribe now for unlimited access to everything PS has to offer.

As a registered user, you can enjoy more PS content every month – for free.

Already have an account? Log in

Secure your copy of PS Quarterly: The Year Ahead 2023

Our annual fourth-quarter magazine is here, and available only to Digital Plus and Premium subscribers. Subscribe to Digital Plus today, and save $15.

Subscribe Now

Laura Tyson

Writing for PS since 2012
112 Commentaries

Laura Tyson, a former chair of the President’s Council of Economic Advisers during the Clinton administration, is a professor at the Haas School of Business at the University of California, Berkeley, and a member of the Board of Advisers at Angeleno Group.

Susan Lund

Writing for PS since 2013
13 Commentaries

Susan Lund is a partner at McKinsey & Company and a leader at the McKinsey Global Institute.

new comment has been posted. new comments have been posted.

0 Comments on this paragraph, 25 Comments on this article

Sean Matthews Dec 13, 2017

I have read the McKinsey report. It is a scaremongering joke that tells me more about the gullibility of McKinsey than about plausible futures. It is not a piece of serious research. it reads like one of those Sunday supplement articles I used to read as a boy about how super the future will be. There is no reason whatsoever to take it seriously - It pluck estimates (accurate to two decimal points) about the percentage of work in various sectors that will get replaced using methods with next to no empirical dimension (not surprising, since there is no empirical evidence with which to calibrate any method), while also acknowledging that there are no proofs of concepts existing of any of these applications, or initial deployments, and there is no evidence of wondrous leaps in productivity in any reports from the actual economy.

They seem to think that deep learning is automation pixie dust.

Speaking as somebody who actually consults on, and builds, advanced machine learning and analytics software for a living.
- Steve Hurst Dec 13, 2017
  
  Sean - the fact you are aware of the real world implementation problems does not mean the technology cannot function perfectly well in repetitive tasks and the focus of scale is to reduce unit cost via consistency and compliance and repetition, ideal preparation for full automation. Foxconn have replaced production line workers with robots, Uber are intent on replacing divers with robos as two examples. Apart from the automobile it is doubtful there has ever been so much collective human effort undertaken as will occur with AI. The Model T bears little relationship to the current automobile. Automobile ownership expanded with development, so will AI. Uber consider something in the region of 40% of their cost is human drivers. I havent been in a car factory lately but I would expect robot bodyshell welders not humans. Anything in a closed or predictable environment is up for target
  - Sean Matthews Dec 13, 2017
    
    I did not say, and do not think, that there will be no incremental productivity improvements in the future. That is not the claim here - the claim here is that there is a looming wave of automation that in a non-incremental way will replace a material proportion of the current workforce. There is no reason (certainly the McKinsey report which functions as the core of this article - if read with the slightest modicum of critical scepticism - presents no convincing reasons) to believe that this is going to happen. And there is no indication in empirical economic data that is is starting to happen.
    - P. Kamath Feb 26, 2018
      
      Sean: I have a serious interest in reading this report. I share some of your views. Would you kind enough to send me/ or links on this PS Please?
      A new reply to this comment has been posted.
    - Steve Hurst Dec 13, 2017
      
      Sean, I'm not accusing you of anything. The Foxconn robo production line was a step discontinuity and they will replicate it if it works, which is highly likely. It depends where you are stood in the food chain. UK retail banks are moving away from a human interface. Dispensing with humans creates other problems that's nothing new - The key thing is a closed system or environment with predictable functions giving automation opportunity. Hotel reception, travel booking or restaurant meal ordering is an obvious to name a few. Empirical data is not a lot of help for something new. I fully agree your implied assessment that implementation is not easy but there is so much opportunity it will occur. For a large implementation it will be step by step (which is the mode in UK retail banking) followed by rapid acceleration. Healthcare (driven by an ageing demographic) is where you are likely to see the most effort in an attempt to contain rapidly expanding demand in a politically sensitive arena. The pressure to do something is going to be immense and the consumer will wear it if provision - which would otherwise be curtailed - is made. You're talking about a couple of decades at the most. The key with AI is the cost of replication is near zero once functioning which breaks through the usual incremental pattern. Physical robotics has a different layer again. The problem for society will be the sudden acceleration of implementation following validating the application. To quote Hemingway as an analogy - “How did you go bankrupt?"
      Two ways. Gradually, then suddenly.” The Sun also rises.
      A new reply to this comment has been posted.
    A new reply to this comment has been posted.
  A new reply to this comment has been posted.
A new reply to this comment has been posted.
Prashant Kotak Dec 12, 2017

This article is either delusional - or (sinisterly) disingenuous.
If the authors believe what they write, I would challenge them to point me to a couple of dozen professions that they think will remain human-only preserves over the next two decades. I am willing to bet that for every profession they come up with, like Priest, Teacher and so on, I can point to at least ten that will become *fully* automated.

"But history and economic theory suggest that fears about technological unemployment, a term coined by John Maynard Keynes nearly a century ago, are misplaced."
This is rot - history says the future carries on as in the past - until it doesn't. As for economic theory, just don't even go there, for fear of economists being laughed out of court. The truth is there we are approaching the end of the road for the type of mass employment that the last industrial revolution created. There is no more road. If you think other types of work will be created that will provide work for billions then I challenge you - point me to those professions - and I will point you right back at those same professions being also automated in short order.

"These trends must be reversed. Lifelong learning needs to become a reality."
Sure, but this will offer a stay of execution only and that for a couple of decades tops. What then?

A final challenge, on the numbers quoted - "75-375 million workers" and other figures quoted are hokum - or if you prefer, economic theory. There is evidence that almost 30% of all work around *right now* is automatable, but hasn't been automated for various reasons, and this is provided as argument that automation will unfold at a linear leisurely pace into the future.
I disagree.
Technology-driven change is like pulling on a very large brick with a rubber-band: for a long time nothing seems to be happening - and then, just when you expect it to start creeping towards you, it instead wizzes through the air - and smashes you in the face. Regarding large-scale Automation, it's already started and we stand, right now, on the edge of a startlingly fast ramp upwards. Once some businesses start to automate, others are forced to follow - or go bust. The speed of change will acclerate as professions like accountancy and law start to lose (very smart) workers - these type of people can and will retrain and move into still-viable IT related professions, speeding up the process. My fear is for the billions of dirt poor people in Africa and Asia. How do these people re-invent themselves? Gene-editing or eugenics?

I'm not saying at all that humanity should go back to the horse and plough; for the start the option doesn't exist without creating ISIS type dystopias. But let no-one peddle the delusion that what cones next is anything other than chaotic, and likely catastrophic for the majority of humanity.
- Steve Hurst Dec 13, 2017
  
  Prashant. I think the problem in Africa is different. 60% of the population are agri workers. Somewhat like 17th or 18th century Europe. Europe now has 2% of the population working on the land. If that happens which is not AI or robots then a huge number of land workers become unemployed. Much the same in India. I think it could be catastrophic but it is never discussed in the big plan to industrialise these locations. As for AI, yes its deployment is already in place and underway and the speed it can be replicated (and updated) is the thing. You dont need to build a factory complex like Henry Ford had to at Rouge River or acquire thousands of acres of mines and rubber plantations
  - Prashant Kotak Dec 17, 2017
    
    Steve, your line of argument helps crystallize the reasoning behind the sense of foreboding I have about Africa and Asia. Africa will industrialise, for a start that is what the Chinese have promised the African elite (chieftans and tribal generals and their descendants) who very much want that.
    
    But this time industrialisation won't provide a century and a half of employment in factories and officies and a living for millions - that stage will be mostly by-passed because of automation. People working on the land as subsistence farmers may decide to stay put instead of migrating to the cities in huge numbers once they see which way the wind is blowing, but equally large-scale industrialised farming is unaviodable in Africa, and also likely pushing many poor off the land will be forced purchases by govenrnments as they build motorways and so on. Also, the population explosion in Africa means millions will have no option but to migrate looking for urban work - subsistence farming is unfeasible below about an acre of land per family.
    
    So the question then becomes, how will those millions survive, what will they do, if the cities provide no work?
    - Steve Hurst Dec 19, 2017
      
      PK, You know what they will do, because it is what you would do
      A new reply to this comment has been posted.
    A new reply to this comment has been posted.
  A new reply to this comment has been posted.
A new reply to this comment has been posted.
Walter Sweet Dec 11, 2017

Even though I'm a progressive, I always enjoy comments talking about retraining workers. I'd suggest there is little evidence these work, save for healthcare and those jobs may increasingly be automated too. While it is true the past suggests productivity gains etc and innovation will create new jobs, it's not clear this will happen in sufficient numbers, even with fancier education/training programs, to prevent instability and suffering and leading to more autocratic regimes.
A new reply to this comment has been posted.
Zijian Ma Dec 11, 2017

This analysis is based on a few shaky assumptions. Can the new era guarantee the advent of new occupations? Can retraining programs effectively ease the transitional pain caused by artificial intelligence, if people cannot even know what to learn?
A new reply to this comment has been posted.
Derryl Hermanutz Dec 11, 2017

In his 1817 opus, Principles of Political Economy and Taxation, in Chapter XXXI On Machinery, that great advocate of comparative advantage and free trade David Ricardo admitted that the Luddites were right: replacing workers with machines “renders the population redundant, and deteriorates the condition of the laborer.”

The displaced workers never found “new jobs” in the now-mechanized economy. The unemployed European masses were “offshored” by the millions to populate 3 ‘empty’ continents: North and South America, and Australia.

There are no more empty continents to accommodate the new 100s of millions of unemployed masses. There are no new continents to “develop”, building up agricultural, industrial, residential and transportation infrastructure. There will be “no jobs” for the new “technologically unemployed” masses, just like there were no jobs for the old unemployed masses.

The idea of creating jobs, while the capitalist-owned productive economy strives to eliminate paid workers to reduce costs and increase profits, is an exercise in the creation of what David Graber calls “bullshit jobs”. Everybody would be better off if all those bullshit workers were paid the same incomes to stay home. They would still be producing nothing of any use to anybody; and they would still be spending their incomes buying all the stuff produced by the capitalist-owned machines. But at least they wouldn’t be contributing to rush hour traffic driving to ‘work’.
- Gene Coyle Dec 23, 2017
  
  Excellent. Why is this not obvious to all?
  A new reply to this comment has been posted.
A new reply to this comment has been posted.
Steve Hurst Dec 10, 2017

Globalisation is about increasing the supply of labour to Capital's advantage. That is why Capital has invested in it. Ever heard of FDI in Asia.
Those increasing the supply of labour are not interested in compensation to those adversely affected. QED all around
AI is about increasing the supply of labour as AI replaces human labour. This again moves the labour supply and demand point further in favour of Capital
Increased automation may create other jobs but depending on the penetration of AI those jobs will be dwarfed by job losses.
In the late 19th century 40+ percent of the population was engaged in agriculture. Due to mechanisation that has dropped to 2%.
Sure, unemployed agri workers got other jobs at the time, but that depended on other sectors growing very significantly. In the early 20th century that was due to the development of consumerism, now a done deal.
It is totally ridiculous to suggest that the growth of new jobs which has been struggling for some time can suddenly and in a matter of a couple of decades fill the gap that will be, and is right now being, created by AI implementation. Equally it is ridiculous to suggest that a measure being implemented to reduce labour will somehow create even more labour. If that was the case it would not be being implemented.
As for compensation to sacked workers, where is that going to come from. The volume of money needed is beyond any simple realignment.
One thing is very certain. When AI comes to developed economies there will be increased unemployment; and when mechanisation comes to developing economies there will be massively increased unemployment.
As for the red herring of retraining newly jobless workers. The big problem with this is employers prefer fresh fruit. Again, QED.
'These marvelous new technologies promise higher productivity' Really? Why has productivity been in decline decade upon decade when new tech has been implemented during that period. This claim is utter nonsense and without any basis whatsoever
A new reply to this comment has been posted.
Steve Hurst Dec 8, 2017

Hopelessly naive
A new reply to this comment has been posted.
B Wilds Dec 8, 2017

Americans don't just want jobs, they want "good jobs" and this means full-time work that pays well and has benefits. Today few occupations are totally secure from the wave of labor-saving technology currently in the pipeline. This means it is the job of those we have sent to Washington to create or paint a vision that shows how America can integrate labor-saving devices into a job-creating bonanza or at least an economy that is sustainable going forward. The fact is in my area while we continue to hear about unemployment at sixteen-year lows many people remain underemployed and are struggling. It should be pointed out that going forward those on the government dole or a guaranteed income may find they are at the mercy of a system where at any time benefits are canceled or cut. The article below delves into this important issue.

http://We Need Good Jobs That Are Sustainable html
A new reply to this comment has been posted.
stephan Edwards Dec 7, 2017

None and I mean NONE of this going to happen and everyone reading this article knows it. You lost your job paying a middle class wages and the only job available wiping arses in a nursing home at minimum wage tough that's your problem. Go back to school by going into 50K worth and hope that you can find someone to hire and they automate your new job or ship it to Mumbai before you pay your loan.
Bluntly isn't going to be nor frankly is there any interest by the politicians or their corporate masters in helping the displaced. Whether displaced by off shoring or by automation doesn't matter not a single solitary thing will be done.
There isn't any money for work force training (We spent it all on a tax cut) and there isn't interests by the countries real owners not their problem.

With all due respect if you the western governments any western governments are going to do a single solitary thing beyond speeches and sound bytes for losers. You are probably the only one!!!!!
A new reply to this comment has been posted.
Jerry hough Dec 7, 2017

The first item on the agenda is a limitation of legal and illegal immigration, family unification, and the ability of those with student and worker visas to stay. Even those who get married should have to go home and apply to immigrate. With English and skills they should do fine. And there needs to be a revision of the 14th amendment--which can be done by Congressional action--and an integration of INS and Social Security data sets with real federal IDs and severe punishment for employers. Of course, there needs to be amnesty or the like for those here. To argue that automation is crucial and then to push for an expanded labor force through immigration is simply shameful. Shameful.

Unfortunately, the rich suburbanies of the North and South control the Democratic and Republican parties respectively. All they want is a surging market (up 32X since 1982) fueled by stagnant wages associated with immigration and outsourcing and economic measures permitted by a common divide-and-conquer policy towards white and minority lower and middle income.

Let us hope this is not Germany of 1928
A new reply to this comment has been posted.

https://prosyn.org/fXuu1WD

Rage Against the Machine?

Secure your copy of PS Quarterly: The Year Ahead 2023

New Comment

Contact us

Handpicked to read next

Europe’s Chance to Lead on Robotics and AI

Log in/Register

Rage Against the Machine?

Secure your copy of PS Quarterly: The Year Ahead 2023

Handpicked to read next

Edit Newsletter Preferences

Set up Notification