Slashdot Mirror
Economists Say Newest AI Technology Destroys More Jobs Than It Creates
HughPickens.com writes: Claire Cain Miller notes at the NY Times that economists long argued that, just as buggy-makers gave way to car factories, technology used to create as many jobs as it destroyed. But now there is deep uncertainty about whether the pattern will continue, as two trends are interacting. First, artificial intelligence has become vastly more sophisticated in a short time, with machines now able to learn, not just follow programmed instructions, and to respond to human language and movement. At the same time, the American work force has gained skills at a slower rate than in the past — and at a slower rate than in many other countries. Self-driving vehicles are an example of the crosscurrents. Autonomous cars could put truck and taxi drivers out of work — or they could enable drivers to be more productive during the time they used to spend driving, which could earn them more money. But for the happier outcome to happen, the drivers would need the skills to do new types of jobs.
When the University of Chicago asked a panel of leading economists about automation, 76 percent agreed that it had not historically decreased employment. But when asked about the more recent past, they were less sanguine. About 33 percent said technology was a central reason that median wages had been stagnant over the past decade, 20 percent said it was not and 29 percent were unsure. Perhaps the most worrisome development is how poorly the job market is already functioning for many workers. More than 16 percent of men between the ages of 25 and 54 are not working, up from 5 percent in the late 1960s; 30 percent of women in this age group are not working, up from 25 percent in the late 1990s. For those who are working, wage growth has been weak, while corporate profits have surged. "We're going to enter a world in which there's more wealth and less need to work," says Erik Brynjolfsson. "That should be good news. But if we just put it on autopilot, there's no guarantee this will work out."
When the University of Chicago asked a panel of leading economists about automation, 76 percent agreed that it had not historically decreased employment. But when asked about the more recent past, they were less sanguine. About 33 percent said technology was a central reason that median wages had been stagnant over the past decade, 20 percent said it was not and 29 percent were unsure. Perhaps the most worrisome development is how poorly the job market is already functioning for many workers. More than 16 percent of men between the ages of 25 and 54 are not working, up from 5 percent in the late 1960s; 30 percent of women in this age group are not working, up from 25 percent in the late 1990s. For those who are working, wage growth has been weak, while corporate profits have surged. "We're going to enter a world in which there's more wealth and less need to work," says Erik Brynjolfsson. "That should be good news. But if we just put it on autopilot, there's no guarantee this will work out."
If the job still gets done it's a good thing that jobs gets replaced by AI.
The flaw isn't in who does the work, but how the economic system around it is set up.
That's what jobs used to be, work, stuff that you don't want to do, hence getting paid to do that stuff. Modern technology is invented by people who think: "That looks boring, dangerous and/or unhealthy. Let's find a way to get rid of that work." Destroying "jobs" is the very purpose of technology. If people find work that was previously unnecessary, then that's essentially a negative side effect (although usually combined with the positive side effect of a higher standard of living through higher total productivity). But still, "creating jobs" has never been the purpose of technology.
Its not about the Jobs, I would be glad without a Job - I just need the money. Why? Because I have to pay someone to do things I cant/wont.
But if its robots all the way down, who should I pay? The man who owns the robot? Well I would but I have no job. So we can all agree that we have all things for free since robots made them and the robots get all the stuff to make robots and so on (robots all the way down) or we have to create bullshit jobs no one needs to distribute the money, till someone finds out that we don't have to if we just give things away for free because there is no one who needs to work anyhow.
The only reason the robot exists, is because the man who owns it paid someone to build it for him, or if he built it himself, paid someone for the components. He would only do this if he expects a return on his investment. I assume, that for him, a robot would be cheaper than paying for a human to do the work. So, he would be able to make more profit. > So we can all agree that we have all things for free since robots made them No, the man who owns the bot wont let that happen. >or we have to create bullshit jobs no one needs to distribute the money No one is going to pay anyone for doing a bullshit jobs. The only way out of this problem, is if everyone gets paid a Basic Income by the government. Money for nothing. Its inevitable this will have to happen.
Take self driving cars for example. Once they're good enough to be on the road safely, insurance companies will notice that their accident statistics are lower than human drivers. So first of all they'll lower the insurance for them. Somewhat later they'll put up insurance for human drivers. Then after that some companies will refuse insurance for any manually driven car. Then they all will. And not long after that governments will ban human driven vehicles entirely from public roads.. I reckon this time frame will be about 30-50 years.
Now this might come as a surprise to some of the technokids out there - but some of us actually *like* driving and don't want a computer doing it for us.
As fas as building AI goes, this famous quote is very valid - just because they can doesn't mean they should.
Now this might come as a surprise to some of the technokids out there - but some of us actually *like* driving and don't want a computer doing it for us.
Well... The public roads aren't for joy riding. It's infrastructure for transportation. One might very well argue that you do not have the right to subject other people to unnecessary risk, just because you want to have fun.
Luckily the US has plenty of desert and car-crazy people, so if public roads were closed to human drivers, I'm sure there'll be lots of race tracks and open areas were human drivers are still allowed, etc...
Why should public roads be a government subsidized joy ride arena?
This isn't an overpopulation problem. It's a "we still expect people to pay for food and shelter even though we don't need anyone to do any work" problem. AI is going to force us to grow out of capitalism.
Economists are finally getting concerned because AI can replace them.
when robots came for maids,i didn't cry out as I wasn't a maid
when robots came for factory workers, i didn't cry out as I wasn't a factory worker.
When computers came for book keepers, i didn't cry out as I wasn't a book keeper
now the machines are replacing politicians and lawyers and I cry all the time but no one tries to help me.
i thought once I was found, but it was only a dream.
It's not clear that Apple could survive in isolation. A lot of their components are only as cheap as they are because of other lower-margin companies paying a big chunk of the R&D costs. When Apple was using PowerPC processors and were the only customer for IBM or Motorola for a particular chip, they found it very difficult to compete. They're designing their own ARM cores now, but they're benefitting enormously from the thriving ARM software ecosystem.
I am TheRaven on Soylent News
Now, I'm no optimist on the imminent-coming-of-strong-AI; but this I do know: The University of Chicago does not specialize in producing lefty-pinko-economists. They have departments with a much stronger liberal bent; but econ sure as hell isn't one of them. It's pretty much the altar of Milton Friedman, the school that made the 'Chicago boys' of Latin American, um, repute. If the UofC says that robots are screwing the proletariat, I'm going to err on the side of caution and suspect that the proletariat is screwed...
"you don't want to destroy the redundant people, they're what really makes your economy."
Please, apply a bit more of imagination.
*Current* economy, not much more than a century old (since Henry Ford, to put an obvious time tag) is based on a middle class buying production.
But for basically all history, wealth distribution has managed to work on a basis of a very short affluent/powerful class with a majority of peasants/slaves/outclassed. Maybe the 20th century has just been an exception along history and we are just returning to the standard trend.
Think again digital janitor scum.
Sincerely,
The 1%
Humans are be able to do repetitive intellectual work. This is starting to be automated away.
Its really not, we've made zero progress in actually making machines that can act intelligently and creatively. We can make at best imitations that try to fool one into thinking that there is creativity, and we can use brute-force searches on certain types of problems. Actual innovation is not something we have seen, nor (IMO) will we ever see from AI-- and certainly not until we make phenomenal bounds in understanding consciousness.
About 33 percent said technology was a central reason that median wages had been stagnant over the past decade, 20 percent said it was not and 29 percent were unsure.
Which means nobody has any real idea and the data isn't conclusive yet one way or the other. Furthermore economists are noted for being unable to come to a consensus. There's an old joke that if you ask 10 economists about something you'll get 11 opinions. If they do come to a consensus about something THAT is worth paying attention to. Otherwise it is pretty much business as usual. I also think that you'll find that those percentages correlate heavily with the political leanings of the economists being polled in this very unscientific poll.
More than 16 percent of men between the ages of 25 and 54 are not working, up from 5 percent in the late 1960s; 30 percent of women in this age group are not working, up from 25 percent in the late 1990s.
Umm, perhaps that has quite a bit to do with the fact that we're still recovering from the Great Recession. You know, the economic problems of the last several years that have NOTHING to do with AI or automation and EVERYTHING to do with finance run amok? Hell, prior to the crash in 2008-9 unemployment was at historic lows.
A few years back there was a great deal of interest in computers doing visual processing and recognition, and I was doing a little work in this area. The interest is still there, but news about it seems to have retreated from the front page. The security industry was especially interested in facial recognition. Alongside that interest were the usual peddlers of hype and hysteria. It was difficult to sort through all the noise. When I looked into research papers, I found that the details told of all kinds of limitations. Yes, they could match faces with 90% accuracy. If the lighting was good. And was the same level in the two photographs. And the subjects were all facing the camera at the exact same angle. And the subjects hadn't grown or removed any facial hair or glasses, or even changed hair styles. And they didn't have different expressions. And the database didn't have more than a few hundred subjects. But never mind, soon we would have video cameras on every street corner, matching every passing face to enforcers' databases of millions of criminals.
Despite the noise, which might lead a cynic to think that it's all hype, facial recognition has improved over the years. It will be the same in robotics. We won't see Robot Basketball Player replace Kobe Bryant anytime soon, no Harlem Globetrotters on Gilligan's Island. But we will see more and better robotics. John Henry scored a pyrrhic victory against a steam hammer. Fighting like that to keep jobs from being taken over by robots is just as useless and futile.
We may yet see that promise of more leisure time come true at last, thanks to robotics. So far, all our labor saving advances somehow have failed to free up much leisure time. Instead, we've put that time towards doing more work. Our parents worked hard so that we can have a better life, meaning, less hardhsip and more leisure time. But it seems more leisure time doesn't automatically make for a more satisfying, better life. Asimov's combination of his Foundation and Robots books had this idea of robots doing so much for us that we became slack and unable to do much for ourselves, and at the same time very unhappy that the struggle had been removed from life to such an extent that it felt empty and meaningless, so that finally we had to abandon the robots. I don;t think that will happen either.
Intellectual Property is a monopolistic, selfish, and defective concept. It is "tyranny over the mind of man"
This is what a universal basic income or citizen's dividend is for. Consumers are the big movers in economy, and producers are the big makers; a portion of all income (individual and corporate) is taken and divided up among everyone (for some definition of "everyone"), stabilizing the bottom.
Imagine if all the homeless and unemployed had a fixed amount of income. Maybe $500-$600/mo. At $1.33/sqft (significantly more than I last rented), a livable microunit housing for a single individual would leave just barely enough for food, utilities, clothing, etc. Just barely. I think I have $50 of leeway in there at 17% of corporate and individual income (eliminating about half of taxes, including OASDI payroll tax, and applying a 17% flat to replace it). Right now, they have nothing, so can't buy anything; in this scenario, they have just enough to buy what they need to live.
This hypothetical creates an enormous market: if you fall to the bottom and lose everything, you still have the shirt on your back, enough money for a new shirt, and enough to rent a sardine can to live in (224sqft microunit; I may be able to get fancy without appreciably increasing costs, too...). Businesses can profit off this, while the mental and physical health problems of being homeless and hungry--starvation, unsanitary conditions, etc.--are lifted off the back of society.
On top of that, producers who fully automate are collecting profits. Automation reduces labor: it costs less to maintain a robot because it takes a collective 10,000 man-hours to produce a robot and 1,000 man-hours per year to fuel and maintain it (including mining fuel, refining fuel, shipping fuel, generating power, transmitting power, maintaining the power infrastructure, mining all the steel for the robot parts, refining steel, shaping steel, and sending maintenance people), but the robot does 50,000 man-hours of work in 10 years. 20,000 is less than 50,000, so that's 40% as many employment hours--40% as many jobs, if you will--for the same useful production.
This labor reduction by efficiency improvements includes far more than automation; for example, Toyota saved 45 seconds from a 65-second process building seats by using a shorter hose (raises the steam temperature) and installing the bolts in a different order (easier, faster access by the tech, who installs bolts and then steams the seats to drive out volatile manufacture chemicals). Many such optimizations allow the same humans to use the same tools to build the same things, but in 80% of the time overall, or 60%, or 40%; thus you only need half as many humans to build as many things in as much time.
The reduction of laborers and the increase in productive output means goods can come cheaper, but consumers are poorer. Fewer consumers exist. A citizen's dividend doesn't free us from work; it leaves us poor, but alive. It frees us from the terrible economic crash that comes when new management styles and processes. We will always find new use for laborers; but this comes after we put laborers out of jobs for a good while, and in the process destroy the labor force. Providing some return to the consumers for being consumes is, thus, advantageous to businesses: it provides them target markets to invest in, avoiding the economic problems of making higher-end goods for the working class which has just become the unemployed, and suddenly not having anyone to sell anything to.
A universal basic income, or a Citizen's Dividend in particular, is the solution to this conundrum. Universal vocational education--that is, college education--touted as a solution, is an exacerbating problem: laborers pay higher taxes or take on enormous debt to flood the market with cheap, skilled labor, giving employers the advantage of lower salaries as unemployment for a skilled labor class increases. Welfare, as a qualified service, takes on more operating cost as more people collect; while a universal income always pays at 100%, and is thus immune to the fluctuations of economy.
Support my political activism on Patreon.
You can't mine iron much faster with more information at hand, crop yields don't increase with more information at hand. Travel times aren't reduced since several decades, and where they are indeed reduced...
Totally disagree. Not to nitpick words, but information by definition is useful data that you can understand & incorporate. So unless you got a ton of just raw useless data (ie: just a simple text file of first & last names of every person who went on site) on your We Mine Iron Inc. server then the information would certainly speed up your mining to consumer operations. Same with crop yields. Information is just as useful and many times more so as mechanical efficiencies. If one doesn't see the increase in productivity, then they don't really have useful data (no information) or they don't know how to properly measure it.
Travel times? Work from home. Video conferencing. Remote monitoring. Smart Grids. Smart Factories. Parking Reservation Systems. Online Shopping. Video Funerals! Another way to look at it is that travel time has been reduced from days & hours to 5 minutes.