Will Machines Render Low-Skilled Workers Useless?
- Posted by ryan on August 9th, 2009 filed in Miscellany
No. Greg Clark suggests otherwise, noting that computers and other machines are increasingly capable of doing things that only humans could previously do. The ranks of the unemployable will only grow over time, he says, potentially leading to a world in which the bulk of the population sits around idle, supported by the few skilled workers who are still smart enough that machines can’t duplicate their work.
But this is silly. Why? Machine and robotic resources aren’t free; they’re resource constrained just like everything else is resource constrained. We have the tecnological know-how to replace millions of human workers with machines right now, but we don’t because the expense of building, programming, operating, and maintaining the machines is too great. It’s not worth it. As demand for human labour falls, the price of human labour will also fall making the hiring of humans more attractive. Meanwhile, as demand for robot labour increases, the price of robot labour will also increase (since the stuff robots are made of is scarce), making the use of a robot for any given task less attractive. There will then be some market equilibrium which will, in all likelihood, involve plenty of employment for low skilled workers.
Which doesn’t mean that machines can’t place downward pressure on unskilled worker wages. They clearly can; the rise of the computer destroyed the jobs of millions of semi-skilled clerks. Those people are still working today, generally competing with low skilled workers and earning less than they previously would have. On the other hand, the rise of computers has enabled them to get a lot more utility out of what they do earn.
There are multiple constraints limiting the total dominance of machine labour. One is the analytical constraint — there are some human cognitive and physical processes that we haven’t yet learned to emulated in machines. This will continue to be less binding over time, enabling workers to potentially compete with humans in a steadily broader range of fields.
Then there is the energy constraint. Machines require power to operate, and the more machines we build, the more power they’ll need. This constraint might eventually be overcome, but until then energy costs will rise with the machine share of the labour force, helping to keep humans at work.
And then there is the materials constraint. So long as it costs money to build things out of other things, it will be difficult and costly to scale up the machine share of the labour force. So until we have superintelligent computers building new machines out of the inexhaustible ether and powering them with a limitless energy source, I think there will be plenty of work for people, though where and how they are employed will probably change significantly.
One other thing: it’s interesting to me that Clark seems to assume that the unskilled will bear the brunt of the pain from machine competition. It seems probable to me that machines will begin replacing doctors before machines begin replacing, say, streetsweepers. It’s not inconceivable to me that the increasing facility of computer programmers may ultimately lead to a diminishment of income gaps rather than a divergence, particularly since the resource constraint would seem to be much more important to consider when building a robot construction worker than when building a diagnostic computer that monitors your vitals and essentially displaces your GP.
August 9th, 2009 at 5:49 pm
>> It seems probable to me that machines will begin replacing doctors before machines begin replacing, say, streetsweepers. <<
Sorry, no. Not in the US.
Doctors have complete and deeply entrenched institutional control of the heathcare system.
If and when computer software threatens to replace doctors, doctors will simply have laws passed rendering that software illegal.
All for the good of the patients, of course.
August 10th, 2009 at 2:47 am
I think you’re overstating one case while ignoring a much more important one. Machines or robots or computers might cost a lot up front, but their operating costs are probbly tiny compared to the salaries of a human workforce. I don’t think energy bills or the price of precious metals is going to stop the march of technology.
That said, the fallacy in this argument is the assumption that there’s some finite amount of work in this world, and to the extent that machines do more, there must be less work to go around for the rest of us. I think the best counterexample is in this tour of Craig Venter’s lab: http://www.youtube.com/watch?v=3E25jgPgmzk
It’s very, very long, but it’s interesting. Venter gave genome sequencing a kick in the ass, basically by making it all automated. He went toe to toe with the $5 billion public Human Genome Project and embarrassed them. And that’s been his MO ever since. If you watch the whole video, you’ll see a ridiculous number of instruments and computers as he goes through the lab, but less than a half-dozen people. At one point he points to a table piled high with box after box of box of samples to be run, and mentions that it would have taken a large lab team years to get throught it the old way. Now it’s some absurdly short time. That’s good for biology. But what about biologists? It’s not as if people are avoiding the field because there are these newfangled machines. It just means fewer of them have to waste their time getting their PhDs by stuffing samples into trays. Now there’s more time and money to be spent on other projects.
Now, that’s the cutting edge, and molecular biologists are skilled
workers - they too can see their work automated and their old jobs replaced. But that just frees them up to work on other things. Maybe they’re more able to adapt and find a new project to work on, but I doubt it (science is specialized, and any skilled labor is necessarily less fungible than unskilled work). Either way, it’s just not true that the more work that’s automated, the less there is to go around. If that were the case, we ought to have seen a looser, not a tighter job market in the nineties, as the IT-productivity revolution kicked in.
August 10th, 2009 at 12:15 pm
“. . . the price of robot labour will also increase (since the stuff robots are made of is scarce), making the use of a robot for any given task less attractive.”
I don’t think this is correct. The danger posed by machines is that they are becoming more sophisticated. Sophistication doesn’t follow the laws of supply and demand — just look at computer RAM, which has been getting cheaper at a geometric rate for quite a while. Note that as RAM (and computers in general) grow more sophisticated at a geometric rate, their power consumption rates and the resources required for their manufacture haven’t been increasing much, if at all.
August 10th, 2009 at 2:25 pm
The operating cost of complex mechanical devices are not necessarily trivial. While a vacuum cleaner or electric toothbrush represent relatively mundane and prosaic technologies, airplanes and assembly line robots are at the other end of the complexity spectrum with correspondingly higher operating costs and with the need to perform maintenance based on hours of operation.
August 10th, 2009 at 3:03 pm
There are already examples of human labor being so cheap that it is chosen over relatively cheap machinery. An example is that of the human labor used in ship breaking.
Take a look a Edward Burtynsky’s ship breaking photo series. There you’ll see humans pull large ships onto the beach with steel cables and their hands. Why not use something as simple as a winch? Because human labor is less expensive than the gas needed to run the winch.
August 10th, 2009 at 6:54 pm
One argument you don’t consider here is that once robots become sophisticated enough to potentially replace humans in all forms of industrial labor, then that will include all the steps in the manufacturing processes needed to build more robots–mining, smelting, building and assembling parts, etc. And when (and if) we have full machine self-replication, you can build a single robot factory and as long as it has the necessary raw materials and energy, it can build more robot factories, so the numbers can increase exponentially–wouldn’t this tend to drive the price of any products produced by such factories (including the manufacturing robots that make up the factories themselves) down close to the price of the raw materials and energy?
As for the price of materials and energy, once you have the technology for self-replicating robot factories, it might also be possible to build fully automated asteroid miners running on solar energy (which is about 10 times more powerful in space than on Earth, see here) whose numbers will also increase exponentially once the first one is built, although this depends on whether all the necessary metals for self-replication can be found on a single asteroid (or other body like the moon). If this could work, though, it could cause the price of raw materials to drop a lot–the wikipedia article notes that a single large asteroid has enough iron-nickel ore to “supply the 2004 world production requirement for several million years”.
Also see the wikipedia article on the notion of a post scarcity economy, based on ideas like the ones above.
August 11th, 2009 at 3:35 am
When technology has advanced far in this field, the robotic worker will be tireless, won’t require breaks, won’t get sick’ and doesn’t complain to a workers union. At some point these pros and more will out weigh the costs of buying robots.
August 12th, 2009 at 11:43 am
There may be an infinite amount of work in the aggregate, but a lot of the manufacturing work tends to cross borders. There’s no reason to assume that just because a company automates procedures, the work will not eventually go to lower cost nations.
Somehow, converting a large percentage of a population from manufacturing workers into “symbolic analysts” just doesn’t seem practical.
August 16th, 2009 at 6:19 am
Let me say, for example, that it’s much more difficult to design a machine that can pick delicate fruits (such as cherries, strawberries and tomatoes) or care for newborn infants and severely handicapped adults than it is to design a machine that can make decisions and solve problems at top levels of business or accurately diagnose illnesses and devise the best available treatment plans. So if anything, Prof. Clark has got it all backwards in assuming that it’s easy to replace lower-skilled workers with machines, leaving most of the highly skilled work in the hands of humans.
Look at this way: being a top-notch exec is very much akin to being a grandmaster chess player. Or being a great medical diagnostician is simply someone who’s really good at algorithmic thinking. Needless to say, these are the very things that computers are very good at doing. OTOH, picking delicate fruits or caring for our neediest and most helpless members of society requires a lot of touchy-feely skills that robots can’t even begin to do. In fact, you’d have no trouble finding machines in most of out IT companies with decision making/problem solving skills that rival that of our most highly-prized corporate execs or our most gifted medical diagnostician. By contrast, you’d be hard pressed to walk into the most advanced robotics lab in the world and find a single robot that has mastered the touchy-feely work that’s required of our most highly-prized migrant workers or our most gifted nursing care aides. So for this reason and this reason alone I propose that we replace all of our multi-million-dollar salary execs and our overpaid medical diagnosticians with robots that’ll work for free (24-7), thus enabling us to free up funds for our so-called lower-skilled workers providing them with a decent standard of living!
Let me also point out that as long as commodities, such as crude oil and industrial metals, remain relatively cheap, robots will continue to be a cheap source of labor. But as soon as these commodities are no longer abundant, the cost of building and maintaining robots will outstrip the cost of feeding and housing human labor, making robots not the bargain that they once were. And as the human population continues to grow by leaps and bounds with no end in sight, the commodities that go into the making and maintaining robots will become so damn scarce that we’ll have no other choice but to return to a time when men and mules were the only workhorses of the world. Now if this isn’t dystopian thinking, I don’t know what is…;~)
November 19th, 2009 at 1:04 am
Overpaid diagnosticians? Really? What are you smoking? Those people work hard. Do you always complain when people have more than you because they deserve it?