Summary: Employment growth has been the slowest of any post-WW2 recovery. But a greater problem lies ahead — the robot revolution. The next wave of automation, affecting both manufacturing and the far larger population of service workers. Here’s a status report on its early phase, already in motion. At the end are links to other chapters.
Contents
- Another person discovers the Robot Revolution
- Examples of automation coming now
- The challenge for every developed nation
- Other posts about the Robot Revolution
- Actroid Sara
The animation above comes from “The Robot Revolution: Your Job May Be Next“, The Fiscal Times, 12 July 2011.
(1) Another person discovers the Robot Revolution
Another person discovers the coming next wave of automation: Barry Ritholtz at The Big Picture:
As the chart {of labor participation} above shows, this peaked in 1999, and has been trending downwards ever since. There are several reasons why this is:
- Demographics of the aging baby boomers, who are retiring, living longer, and impacting this ratio;
- Technology/DotCom collapse eliminated lots of malinvestment driven Tech jobs;
- Financial/Credit crash eliminated lots of malinvestment driven banking/real estate jobs;
- Ongoing outsourcing, globalization, etc.
- Robotics
That last item doesn’t get discussed nearly as much as it should, but the single biggest future trend in the labor force is going to be the ongoing replacement of humans by smart machines.
(2) Examples of automation coming soon
.
Today it’s just manufacturing, toll booth and parking lot attendants, and people printing-moving-filing paperwork. But the full dimensions of the revolution slowly become visible.
- MINT Plus, the robot floor cleaner
- AI journalist: “In Case You Wondered, a Real Human Wrote This Column“, New York Times, 10 September 2011
- “Clothing Giant H&M Defends ‘Perfect’ Virtual Models“, ABC News, 6 December 2011
- “Robot prostitutes ‘the future of sex tourism’“,
The Dominion Post, 18 April 2012 - “Robot Professors Come With Singularity U’s Massive Upgrade“, Wired, 22 August 2012 — this will be big!
- “Coming soon: Robots that help build buildings“,
Los Angeles Times, 13 November 2012 - “High-School Video Gamers Match Physicians at Robotic-Surgery Simulation“, Slate, 21 November 2012
The robot above is Baxter: “A unique robot with unique features”, from Rethink Robotics:
Baxter is an entirely new type robot that is redefining the way robots can be used in manufacturing environments. It performs a variety of repetitive production tasks – all while safely and intelligently working next to people. How? Baxter exhibits behavior-based ‘common sense,’ capable of sensing and adapting to its task and its environment. It requires no complex programming or costly integration. And with its uniquely low price point, Baxter provides a compelling alternative to low-cost offshoring for manufacturers of all sizes.
(3) The challenge for every developed nation
Automation increases the return to capital, and decreases the return to labor. It will become a major driver of increased inequality. Only slowly does this realization spread, as in the conclusion from this review in the Financial Times by James Crabtree (red emphasis added):
Yet the author’s more basic conclusion – that technological progress is sufficiently rapid that “many present-day organisations, institutions, policies, and mindsets are not keeping up” – is surely right. The result is a conundrum that shares much in common with trade policy. Technology is essential for creating value and raising productivity, but it creates losers as well as winners. These losers are recognised in theory, but too rarely compensated in practice.
This brings the argument back to a more basic problem: fair distribution. Machines work for free, but their benefits end up in someone’s pocket. If technology is indeed speeding up, more of that benefit must be returned to those it affects, especially in the form of investment in human capital. If not, the march of the machines will overtake us sooner than we think.
(4) Other posts about the coming Robot Revolution
- The coming big increase in structural unemployment,
7 August 2010 - The coming Robotic Nation, 28 August 2010
- The coming of the robots, reshaping our society in ways difficult to foresee, 22 September 2010
- Economists grapple with the first stage of the robot revolution, 23 September 2012
- The Robot Revolution arrives & the world changes, 20 Apr ’12
(5) Actroid Sara
Actroid is a android developed by Osaka University and manufactured by Kokoro Company Ltd. (the animatronics division of Sanrio). In most cases, the robot’s appearance has been modeled after an average young woman of Japanese descent. It can mimic such lifelike functions as blinking, speaking, and breathing. Plus the ability to recognize and process speech and respond in kind. (From Wikipedia)
.
.
.
FM – – – May we repost your robotics article We have had a lot of discussion about human employment displacement by robotics over the past several days as a result of Steven Hansen’s discussion of employment problems in his weekly review article. John
John B. Lounsbury Ph.D. CFPManaging Editor Econintersect.com
Senior Contributor TheStreet.com, highly ranked author Seeking Alpha
One idea to help offset the growing “robotic deployment effect” on jobs is to tax their use very heavily. These taxes would contribute to a series of funds that would offset unemployment of workers they may replace. I do not see this working very smoothly, and being fraught with beaurocratic inefficiency, but I do like the economic synergy of it. It is sort of like retiring humans and having the robots pay into their financial support structure.
On the whole, this would fail to contribute all the necessary funds to support displaced labor workers. If one could note the makeup of current human workers who COULD be reasonably replaced by robots in the next 5-10 years (time enough to retrain those workers) there could be interesting statistics there. Do we have so many jobs that are conducive to robotic replacement? I do not have these numbers and this is just speculation.
Factory owners would balk publically at the heavy tax, but shouldn’t a robot that replaces a human worker pay social security tax? Just throwing that one out there… :)
“tax their use very heavily.”
This can only work in an economic system where capital cannot relocate from one country to another.
Otherwise, industrialists will just set up plants in those places that have lenient tax rules regarding robotic manufacturing.
Good luck introducing rules limiting the circulation of capital (in all its meanings, whether money to invest or actual machines) in the current economic system. Heck, the EU defines freedom of capital movement as one of the fundamental “four freedoms”. As for taxing capital and means of production, I suggest you look at which direction those specific tax rates have been going in the last 35 years. Downwards.
Robotics is yet another shock (besides growing economic inequality, rapid ecological degradation, natural resource depletion and “reduction in resource density” for the lack of a better word) that will severely disrupt our societies within the next generation. Another riot-inducing parameter, as if we needed more of them.
Ray Kurzweil explains in the below linked video that AIs (Artificial Intelligences) will need to be taught in a way similar to children. So, will AIs be taught by people “parents”, or by other AIs at someplace like Singularity University, or both?
“After Words with Ray Kurzweil“, C-SPAN, 20 Nov 2012 [BookTV.org]
Author and inventor Ray Kurzweil talked about his neuroscience research, in which he reverse-engineered the brain in order to assist in building intelligent machines. Mr. Kurzweil talked about this research with Editor of Scientific American Mind, Ingrid Wickelgren.
Ray Kurweil is an ignorant crank who has been consistently wrong in his predictions. “After the meeting I decided to visit to researchers working on the type of technology that people such as Kurzweil consider the steppingstones to the Singularity. Not one of them takes Kurzweil’s own vision of the future seriously. We will not have some sort of cybernetic immortality in the next few decades. The human brain is far too mysterious and computers far too crude for such a union anytime soon, if ever. In fact some scientists regard all this talk of the Singularity as a reckless promise of false hope to the afflicted.” Source: “Can You Live Forever? Maybe Not–But You Can Have Fun Trying,” Carl Zimmer, Scientific American.
Robots are real and the “deep learning” algorithms that let them perform some tasks incorrectly thought to require human intelligence seem to work reasonably well in the real world. We are likely to get machines that perform up to the level of an insect — dragonflies or moths do an excellent job of navigating complex landscapes while avoiding obstacles — or possibly a dog or cat (which can recognize familiar faces and simple familiar objects like a favorite ball or stick). That’s enough to put many millions of people out of work, because jobs like delivery driver or warehouse forklift operator basically require only the amount of intelligence exhibited by a dragonfly or a dog.
Guest makes an excellent point that adding various cockamamey taxes or fees by playing cybernetic pin-the-tail-on-the-donkey with conventional capitalism won’t work. Conventional capitalism is crashing and burning under the onslaught of robots + databases + algorithms. In fact, the combination of databases + various sophisticated algorithms is destroying a lot more jobs a lot faster than robots. Consider that IBM’s Watson is now training with doctors: how many nurses will lose their jobs at hospitals to a machine that runs patients through a preliminary question-and-answer series before the doctor comes into the room? How many radiologists will lose their jobs to neural nets that do a better job than doctors at spotting fuzzy blobs that might be tumors in an X ray? How many engineers will lose their jobs to genetic algorithms that generate weird but superior designs that no human could imagine?
Our cybernetic problems with capitalism will not be solved by the same procedures or the same minds that created them. We can’t just tweak capitalism a little to fix everything: robots + databases + algorithms are a game-changer, and just as the fall of the Berlin wall discredited communism as a workable system, the continuing deployment of robots + databases + algorithms is in the process of discrediting capitalism as a workable system.
TomDispatch’s “Debtpocalypse and the hollowing out of America” and Robert Skidelsky’s article “Inequality is killing capitalism” both make the same point. This process is driven by technology, and unless we can repeal the basic laws of physics, we’re going to have change our socioeconomic system if we don’t want a world full of starving slaves ruled by a handful of trillionaires who own all the machines that produce everything.
“the continuing deployment of robots + databases + algorithms is in the process of discrediting capitalism as a workable system.”
Why? This sounds like Marx explaining why late 19th century industrialization would inevitably destroy western societies. The allocation of the benefits from increased productivity is a social problem, but why does this “discredit capitalism”? I don’t see the conflict.
“we’re going to have change our socioeconomic system if we don’t want a world full of starving slaves ruled by a handful of trillionaires who own all the machines that produce everything.”
I think that exaggerates the situation. Look at Germany and the Scandinavian nations. They seem to be managing this transition well. Lots of lessons there for us.
“They seem to be managing this transition well.”
A bit early to tell, since such transitions are the affair of one or two generations.
I would be careful about Germany: first of all, because its entire socio-economic constitution has been durably impacted by the reunification with the GDR; they are still paying the price. Second, because its politicians deliberately chose capital over labor — Harz IV and wage deflation has led to a permanently stagnating incomes. Qualified German people are coping by emigrating in significant numbers to other countries (tens of thousands migrate to tiny Switzerland each year, for instance).
As for the much vaunted Scandinavian model, the latest info I read about Denmark indicates that it is getting seriously frayed.
guest,
It’s always “too early to tell” when looking at social trends, since they can always change from the past. All we can see is how well Northern European nations have coped with the automation of the past generation. The answer is clear: far better than we have.
“permanently stagnating incomes”
“Permanently” seems too-strong a word to use. Also, flat real incomes is better than 80% of Americans have managed during the past decade or so.
Plus they have avoided the rapid rise of wealth and income inequality that has afflicted America. Plus, for many of those nations, with a higher rate of income growth.
Their nations are not heaven, but the appropriate real-world standard is IMO a relative one: they’ve done better than most other developed nations.
guest,
Can you give us a cite for net emigration of ethnic germans from Germany? the numbers in a quick search I see don’t give that asa subcategory.
Following a liberalization of the laws — and the one of the strongest economies in the EU — Germany now has net inflows of people, but much of the flow both ways appear to be ethnic non-germans.
Like much of Europe, most of their government data does not track ethnicity or race, just country of birth.
No direct calculations, I am afraid. One has to peruse the data from eurostat — see Migration and migrant population statistics — and the statistics do not go beyond 2010. The last figures showed about a net emigration of German citizens from Germany at around 50000-60000 per year, trend increasing.
re: “Ray Kurweil is an ignorant crank who has been consistently wrong in his predictions. ”
Dunno, possible. He is probably more of an entrepreneur (seller of gizmos and books) than a “real” scientist? If so, that makes for some interesting possible observations on the wisdom of elements of the techno-capitalist elites. (sarcasm)
FM says: ““the continuing deployment of robots + databases + algorithms is in the process of discrediting capitalism as a workable system.”
`Why? This sounds like Marx explaining why late 19th century industrialization would inevitably destroy western societies. The allocation of the benefits from increased productivity is a social problem, but why does this “discredit capitalism”? I don’t see the conflict.’
It discredits capitalism because capitalism as Western thinkers have defined it operates according to rules which are now technologically defunct and impossible. Once upon a time, capitalism produced a “rising tide which lifts all boats.” Now, the tide has risen and 80% of the boats sank. This isn’t an allocation problem and it’s not a social issue. It’s a direct result of the fact that more and more automation generates an increasing amount of revenue with a decreasing number of jobs. Look at the big companies of the 2000s — google, for example, They generate 8 billion dollars of revenue per year with only 19,000 employees. Once upon a time, a company like GM or Ford needed hundreds of thousands of employees to generate that kind of revenue. Projecting into the future, we see that soon 5,000 or 2,000 or 1,000 employees will generate this kind of colossal revenue.
That means that at some point in the near future almost all people in America will be superfluous. They’ll have no function. They won’t be needed. Once upon a time the “lump of labor fallacy” meant that displaced workers would always be absorbed into new industries. Today, the lump of labor fallacy has itself become a fallacy due to exponential automation, and that changes the game entirely. Capitalism, as Adam Smith pointed out, was based on the presumption that selfish behavior by each individual resulted in collective benefit for society as a whole. Technology has now changed so that selfish behavior by each individual now results in collective destruction and collapse for society as whole.
Running time backward, we see the reverse: once upon a time, during the millions of years when humans existed as hunter-gatherers, there was no need for money. Ethnological investigations have shown that economic predictions based on markets are uniformly false when applied to hunter-gatherer cultures. People don’t barter, don’t use money, and markets don’t exist — instead, people exchange goods and services by means of rituals and song and collective celebrations. In hunter-gatherer societies, it was equally false that individual selfishness produced collective social benefit as it is today. From that vantage point we can see that only during a narrow period of technological development, from around the period 8000 B.C. to the year 2000, did capitalism exist and operate as defined by economists today. Outside that narrow period, capitalism didn’t exist and didn’t work the way economists describe.
Just as the post-WW II period was an exceptional socioeconomic era never to be repeated for America, in a larger sense the technological “bloom” from the development of agriculture to high-level automation capable of replacing creative human endeavours was also a unique and limited era never to be repeated. In the same way that once oxygen-produced organisms developed on earth, the methane reducing atmosphere of our planet permanently changed and anaerobic organisms became confined to a tiny ecological niche, now that automation has reached a sufficiently high level to replace creative human capabilities, the ecology under which capitalism formerly flourished has now changed permanently and no longer supports capital markets or the other accoutrements of capitalism.
re: humans are superb, but sometimes reluctant imitators
T. More,
Your interesting views seem to be supported by (informed by?) Boyd/Richerson’s work on dual-inheritance theory:
The Evolution of Free Enterprise Values, Peter J. Richerson (Prof of Environmental Science and Policy, U Cal Davis) and Robert Boyd (Prof Anthropology, UCLA), 2005 — Abstract:
If you can make some brief observations on the above article, I would greatly appreciate it.
Thanks!
Other questions:
If most humans are not employed and making money to spend, what will a plutocratic robotic economy consist of? Will the robots be making and spending money? Would such a system necessarily be completely oppressive? Or could the masses spend most of their time doing something other than work while the robots are busy?
Even if most agriculture becomes automated, people would potentially have significant areas of land that they could use for subsistence and “off grid” life. Assuming that books and libraries still exist in some form, subsistence people would presumably be able to begin to fashion primitive industry to make tools that would make subsistence farming efficient. How would the robot culture prevent that kind of existence without resorting to bald terrorism?
Here is the case for why a totalitarian system run by humans (using robots) would not be stable for long: The Evolution of Free Enterprise Values, Peter J. Richerson (Prof of Environmental Science and Policy, U Cal Davis) and Robert Boyd (Prof Anthropology, UCLA), 2005 — excerpt:
Final question: if Darwin (and Richerson/Boyd) are correct about the evolutionary origins of human culture and psychology and the imperfect human tendency toward democratization, justice and equality, why would robotics have to be destructive of the regulating mechanisms wired into human culture to prevent permanent totalitarianism/oppression by plutocrats? Wouldn’t some of the surviving anarchists figure out how to reprogram robots to work against plutocracy?
Economies are based on exchange, which requires markets as a price-setting mechanism. It’s not at all clear that if technology becomes sufficiently advanced that there will be any need for exchange of goods or services. Fred Pohl wrote an amusing sci fi story about this issue in the 1950s called “The Midas Plague.” In that story, robots generated so many goods and services that humans couldn’t consume them all. This made people who failed to consume their required portion of goods and services poor in the Midas economy, while those who consumed more than their share were considered wealthy and achieved greater social status.
We can see the beginnings of a non-exchange economy today with Wikipedia and free open wifi hotspots and public tool lending and open public 3D printer and numerical machine tool cooperatives. These new types of economies are based on “whuffie,” that is to say, collective judgement of value or trustworthiness.
This is the answer to your first question. For a book describing a “whuffie” economy, see sci fi writer Cory Doctorow’s first novel, Down and Out In the Magic Kingdom.
Many types of goods are now so plentiful in advanced countries that they are given away: indeed, failure to get rid of surplus amounts of these goods or services would cause problems because many of the systems that operate in the modern economy are predicted on fulfilling a certain level of production or usage of goods and services, and failure to use this amount creates bureaucratic problems.
For example: most of the surplus food given away to the poor is given away in order to reduce surplus stocks from food production which if not held off the conventional market would cause prices to drop. The surplus bandwidth given away free in open wifi hotspots must be used because the contracts with ISP backbones require a certain minimum level of bandwidth service by the ISP. And so forth.
Marshall Brain has written a book (Manna) about a hypothetical future society in which all the essentials of everyday life are free: food, transportation, access to information, and so on. Robots produce so much of this stuff that it’s treated like free wifi bandwidth today — a public commons, not worth charging for, just as during the MIddle Ages public grazing lands were commons and were not a medium of exchange or subject to barter or markets. Elements of such a society are currently falling in place today, including free food giveaways and free meals in most cities and towns in America (typically done by religious organizations), free tool lending libraries (see this Wikipedia article listing cities that have these). For evidence that hunter-gatherer societies do not use markets or money and do not structure the transfer of good or services by means of exchange (which would require money or markets to set prices), see David Graeber’s essay “On the Invention of Money.”
The answer to your final question is that robots in and of themselves are not necessarily destructive to either an economy or to democracy — which I believe is FM’s point in his above post. What robots will be destructive to is conventional market capitalism as it has existed in the recent past, particularly from the era 1914 (Henry Ford’s invention of the assembly line) to 2000. The socioeconomic structure of advanced societies will have to change as robots continue to replace humans. How this change will occur or what the changes will consist we can’t say precisely, just as the Assyrians who invented acqueducts and mass farming could not foresee modern wheat futures markets as the eventual outcome. But we can see certain broad outlines of likely social changes, as FM has pointed out. Continuing reduction in labor force participation rates, an expansion of goods and services held to be community-based commons, and an enormous increase the amount of open source peer production as the primary means of production. This implies socioeconomic systems based on reputation rather than production, in which consumption is dictated by your collectively judged social standing rather than a price-setting mechanism like a capital market. A number of current sci fi writers are currently exploring future societies that work this way, and a number of social collectives like ebay or wikipedia already work this way. (On ebay, it doesn’t matter how much money you have to spend if you have a poor social rating. Only people with a certain minimum amount of positive feedback can buy items.)
Regarding the Richerson & Boyd paper WTF cited:
Jared Diamond explained why some tribal societies became advanced technological nations while others didn’t in his superb book Guns, Germans and Steel. The basic answer is climate, geography and distribution of native plants and animals. For example, there is no equivalent to the horse in South America which can be tamed and used by riders. This hugely retarded technological progress in South America.
A partial answer to “why now, and why so rapidly” may involve the fact that technological progress involves networking. This clearly isn’t possible with isolated bands of hunter-gatherers because any technological advance typically gets lost when the innovators die. Development of agriculture + animal husbandry allows libraries (the oldest existing Hittite clay tablets are records of grain sales), which in turn exponentiates networking of information. It took a long time for networking to get from clay tablets of grain sales to the internet because informational networks work on an exponential connective system where the number of possible connections twixt nodes scales roughly as the number of possible combinations of paths between nodes, which is well known as a non-polynomial, i.e., exponential growth, system. Viz., the travelling salesman problem, which also describes the growth in number of possible connections in any communication system as the number of nodes increases.
The characteristic of exponential growth system is that they start off extremely slowly and then accelerate at incredible speed. A classic example is the trick question: given a pond in which the number of frogs doubles every day, how many days will it take between the time when there are 10,000 frogs and more than a million frogs? The answer is 7 days, because 2^7 = 128 and 128 * 10,000 is more than a million.
I believe that this is the point FM and the authors he cites, like Marshall Brain, are trying to make. Exponential systems tend to change slowly until they reach the knee of the exponential curve, and then reach a tipping point at which a cascade of massive changes occurs very quickly. The development of robots seems to offer a classic example of such an exponential system.
Pingback: In Friday’s job report you’ll see early signs of the robot revolution! | Automation Impact on Jobs
Pingback: Giornalisti e robot | USIGRai
Pingback: Il redattore che non beve caffè e lavora alla velocità della luce | LSDI
Pingback: Primary Research | Lab A - Blogspot