10

According to this Huffington Post opinion piece, it's possible that in the future the ongoing process of automation will lead to less job opportunities.

It quotes economist Paul Krugman arguing one way:

information technology would end up reducing, not increasing, the demand for highly educated workers, because a lot of what highly educated workers do could actually be replaced by sophisticated information processing -- indeed, replaced more easily than a lot of manual labor.

It also quotes (and tries to refute) economist Brad DeLong arguing another:

I don't see a problem with the number of jobs: I don't see any reason that technological unemployment should be any more in our future than it has been in our past.

Is the automation and increasing productivity leading to job losses?

Oddthinking
  • 140,378
  • 46
  • 548
  • 638
  • 2
    It is difficult to predict the future, and probably impossible to do it reliably (and therefore this question may be unanswerable). My guess is that it will decrease jobs of the type that we know about today. People need to hope or expect that these will be replaced by jobs that we don't know about. My first boss, in the early 1980s at a telco equipment manufacturer, told me, "I don't know what I'll be doing in two year's time: all I know is that it will be a job which doesn't exist yet." – ChrisW Aug 04 '13 at 11:45
  • 5
    Those are exactly the same arguments that have been done over and over again over the course of history when some new innovation has been made. In the Netherlands, workers used to trow their wooden shoes (sabots) in textile looms to break them as they feared that automation would cause loss of jobs. This is the origin of the word *sabotage*, and was happening in the **15th century**. In the 19th century the [Luddites](http://en.wikipedia.org/wiki/Luddite) were doing similar things. So, history is repeating again one would argue... – nico Aug 04 '13 at 12:07
  • @nico I would have cited [Telephone operators](http://en.wikipedia.org/wiki/File:Telephone_operators,_1952.jpg). But the argument in the article has a modern-day twist to it: it claims, that automation replaces not only unskilled jobs but that the increasing levels of what I'd call AI are also now replacing 'skilled' jobs, for example radiologists': resulting in no opportunities for skilled or unskilled workers: and eventually, because no-one is employed, therefore no-one can be an economic consumer, there will be no demand for anything, and "the modern mass-market economy" will not "survive". – ChrisW Aug 04 '13 at 12:40
  • 1
    @ChrisW: well, but economy will adapt (and is adapting) to these new changes. Obviously it does not happen from one day to the other, but it happens eventually. – nico Aug 04 '13 at 13:03
  • 2
    @Derfder, what sort of answer would you consider valid evidence one way or the other? We have two economists arguing. It is a topic books could be written about; what do you hope for us to contribute? – Oddthinking Aug 04 '13 at 13:31
  • 1
    @ChrisW all that happens is that the nature of labour changes, instead of a girl plugging a switchboard she now looks at a computer screen and monitors for problems she then calls a technician to resolve. And there will always be a need for people like that, that AI will maybe take over her job but then she will be monitoring the AI. – jwenting Aug 05 '13 at 04:52
  • @jwenting If you consider only/specifically switchboard operators, that sounds like a naive theory. A counter-theory proposed in the cited article is that fewer people are needed to implement or monitor automation than were needed to do the work in person, resulting in increased unemployment. – ChrisW Aug 05 '13 at 10:19
  • 1
    @ChrisW was mainly pointing out the fallacy that no jobs are created, only ever lost, where there is automation. With more efficiency, productivity per person goes up, so unless you can claim that total GDP never rises (which means fewer people are needed to produce that GDP) you get the same number of people employed, just each of them doing more productive things, leading to each producing more per hour than (s)he would without automation be able to produce. Of course some fall by the wayside, namely those unwilling to or incapable of learning new skills. – jwenting Aug 05 '13 at 10:24
  • @jwenting But all of this should happen in chat or in an answer. –  Aug 05 '13 at 18:19
  • I think this is an on-topic question. Economists have studied this extensively, and despite their disagreement, we should be able to lay out the evidence on each side of the argument. Ill make my answer community wiki. –  Aug 05 '13 at 21:28
  • regarding deleting question, read this: http://meta.stackexchange.com/questions/96670/deleting-my-own-question – vartec Aug 05 '13 at 23:21
  • The OP requested (through various inappropriate means) that this question be deleted as it was off-topic. It appears to be on-topic. His name has been disassociated from it. (See vartec's [link](http://meta.stackexchange.com/questions/96670/deleting-my-own-question) regarding Creative Commons license. – Oddthinking Aug 06 '13 at 00:45
  • 1
    We need to resolve how this question can best be answered. As @Sancho explains, this has been researched by economists and they continue to disagree. (Is this a real disagreement or manufactured controversy? I do not know.) I invite economists to edit the question to explain how such consensus is resolved; what is the economic equivalent of a meta-analysis? – Oddthinking Aug 06 '13 at 00:48
  • @Oddthinking: The problem with the article is that no one credible disagrees with DeLong. Even Krugman himself considers the logic employed by Mr. ford [to be fallacious](http://www.nytimes.com/2003/10/07/opinion/lumps-of-labor.html). In fact, while Ford cites Krugman, at no point does he say anything contradicting DeLong. Decreasing demand for highly educated workers does not mean decreasing demand for workers (i.e., higher unemployment). It just means there will be less labor incentives to get a degree. – Borror0 Aug 06 '13 at 18:46
  • [The Robots, AI, and Unemployment Anti-FAQ](http://lesswrong.com/lw/hh4/the_robots_ai_and_unemployment_antifaq/) from Yudkowsky's LessWrong. If anyone will take HuffPost's opinion over Yudkovsky's on ANY topic, much less automation, they clearly don't belong on Skeptics. – user5341 Aug 06 '13 at 18:54

1 Answers1

2

Looking forward from 1983

From (Leontief, Wassily; Duchin, Faye. The Impacts of Automation on Employment, 1963-2000. Final Report, 1984):

[T]he intensive use of automation over the next 20 years will make it possible to conserve about 10 percent of the labor that would have been required to produce the same bills of goods in the absence of increased automation. The impact of automation is specific to different types of work and will involve a significant increase in professionals as a proportion of the labor force and a steep decline in the relative number of clerical workers. Because the direct displacement of production workers by specific items of automated equipment will, at least in the initial stages, be offset by increased investment demand for capital goods, production workers can be expected to maintain their share of the labor force.

Looking back

From (Kile, F. (2013). Artificial intelligence and society: a furtive transformation. AI & society, 28(1), 107-115.):

A steady decline of less-skilled jobs began with automation. Employment opportunities gradually declined around the globe.

Two recent agricultural developments illustrate this point: (1) One manufacturer announced development of a driverless tractor. (2) Some farms have attached a chip to their cows. The chip monitors milk production, quality, etc. The cows remain in a pasture until they feel the urge to be milked. At that point, each cow goes to a gate which then opens and from that point on, every aspect of milking is automated.

As computing advanced, the need for workers per unit of output declined, though initially output grew fast enough to mask this phenomenon.

From (Autor, D. (2010). The polarization of job opportunities in the US labor market: Implications for employment and earnings. Center for American Progress and The Hamilton Project):

The job opportunities available to males displaced from manufacturing jobs, particularly those displaced at midcareer, are likely to be primarily found in lower-paying service occupations. While these job losses may be primarily attributable to automation of routine production work and growing international competition in manufactured goods rather than to de-unionization per se, the magnitude of the income losses for males is surely magnified by the fact that the job losses are in union-intensive industries.

From (Freeman, R. (2012). Non-nano effects of nanotechnology on the economy. Nanotechnology: Societal Implications—Individual Perspectives, 68.):

For the past 50 years or so, the displacement effect of technological change has exceeded the employment-increasing effect of expansion of production. Employment declined in manufacturing and agriculture, where technological change is most rapid, and shifted toward services, where technological change is modest.

From (Hayes, B. (2009). Automation on the job. Am Sci, 97(1), 10.):

As for economic consequences, worries about unemployment have certainly not gone away—not with job losses in the current recession approaching 2 million workers in the U.S. alone. But recent job losses are commonly attributed to causes other than automation, such as competition from overseas or a roller-coaster financial system. In any case, the vision of a world where machines do all the work and people stand idly by has simply not come to pass.

Predictions about the future

There are too many predictions to list, but I'll pick a few from just one of the above references:

  • We’ll automate medicine. I don’t mean robot surgeons, although they’re in the works too. What I have in mind is Internet-enabled, do-it-yourself diagnostics. (Hayes, 2009)
  • We'll automate driving. (Hayes, 2009)
  • We'll automate warfare. (Hayes, 2009)
Community
  • 1
  • 1
    http://lesswrong.com/lw/hh4/the_robots_ai_and_unemployment_antifaq/ – shieldfoss Aug 04 '13 at 18:06
  • @medivh Are you suggesting an improvement to my answer? I don't understand what the link you provided is for. –  Aug 04 '13 at 18:08
  • 2
    @Sancho The link says it doesn't lead to unemployment because there's always a demand for labor. Like when 95% of us was employed as farmers to 2% of us now, it doesn't mean there's a 93% unemployment rate. – Kit Sunde Aug 04 '13 at 22:36
  • @KitSunde But, that should be added as an alternative answer, then, not a comment to my answer. I don't see what my answer is supposed to take from the link. I didn't say that "95% of us was employed as farmers to 2% of us now means there's a 93% unemployment rate.". I just quoted studies that said things like "the displacement effect of technological change has exceeded the employment-increasing effect of expansion of production", and "As computing advanced, the need for workers per unit of output declined, though initially output grew fast enough to mask this phenomenon." –  Aug 04 '13 at 22:54
  • 1
    Sancho: I think the problem here (not with your answer, but with the question) is you are quoting the economists on one side, and @medivh and the OP quoted economists on the other side, leaving the economically agnostic unable to see why the models in this answer are more accurate. – Oddthinking Aug 05 '13 at 00:08
  • @Sancho: It's a comment on your answer rather than a full answer of its own because it would make for a poor answer on its own (An answer that is just a single link? Wut?) and it disagrees with your position. You can incorporate it into your answer or not, that is not my call to make, but readers of your answer should see the other side of the argument. – shieldfoss Aug 05 '13 at 06:31