How would humans adapt if robots take all the high-paying jobs, leaving only low-wage jobs?

Question

The question "How would humans adapt if low-wage labor was done by robots?" postulates a world where low-paying jobs are automated away while jobs that require "ingenuity" (i.e, higher-paying jobs) are not:

Background: Nearly all low-wage labor is done by humanoid robots (retail, factories, transport, sanitation, construction, mining, and similar fields). These robots have all the physical capabilities of humans and more, but jobs that require ingenuity and adaptation such as design, management, etc are still done by humans, and technology allowing robots to fill these jobs is millennia away.

Consider an alternative scenario, one that is slightly less plausible, but far more interesting:

Humanoid robots are expensive. This means that they are only useful in the long-term if they can get rid of humans who happen to be even more expensive. They might not be better than the humans, but if they're "good enough", the cost savings would be enormous. So let's automate away all those high-paying jobs, using randomness as a good approximation for ingenuity. Use neural networks to design items, use machine learning and pattern recognition to decide what employees to hire and what sectors to invest in, etc. Get rid of all those silly white-collar "design, management, etc." jobs.

Meanwhile, what about the low-paying jobs? Okay, it would be theoretically possible to automate away retail/factories/transport/sanitation/construction/mining/etc. ...but why? The human employees are already cheap. Any cost savings would be minimal, and it'd just be more efficient to hire "meat" than it is to use machine labor (which would be better suited for all those jobs that require ingenuity and adaption anyway). So, don't bother automating those jobs away.

Now, you have a bunch of extremely smart white-collar workers who are now unemployed. What will they do? Will they accept lower-living standards if they take the low-paying jobs that are still available (and thus compete against the blue-collar workers who will be upset at the new competition)? Will they rebel against the machines who just took their jobs? Will they try to turn the 'low-paying jobs' into higher-paying jobs (and in doing so, risk having corporations trying to automate those jobs away to get cost savings)?

Answer 1

It wouldn't really work, because the decision to do so is in the hands of these high-paid persons that would be left unemployed. Also, in big companies, management decisions on high levels are worth far more than the wage of the manager who makes them. Thus, your claim that

They might not be better than the humans, but if they're "good enough", the cost savings would be enormous

simply isn't true. Or rather it is the lesser part of truth - the major part is that, whilst savings would be enormous, the losses would be even bigger. Only when computers are proven to be way better than humans, especially in crisis and other unplanned, atypical situations, anyone would really consider change.

But let's assume computers did this.

Humans will not get poorer. Computers do not earn money. So where the money went? Nowhere. Still in the market, still circulating, or still being hoarded. The same amount of money, the same amount of people, on average - nothing changes. Possibly the rich would get richer, and poor ones poorer. Or goods may get cheaper and the same wage would allow people to buy more? Hard to tell, because overall it's hellishly complicated set of dependencies.

One way to prevent social unrest would be to make lives of most people better, when firing some of them. On the other hand, in USA white collar jobs are 60% of all jobs. I bet it's similar for other "advanced" countries, like Japan, Canada, most of European Union... No government would ever allow leaving most of the nation without jobs. There would be riots, that's for sure. Army and national guard would be helpless, because every soldier would have a hungry unemployed relative, and they would sympathize. And rulers know this. So you can't just replace white-collars. Not in the countries that are first to be able to afford machined advanced enough.

You need to be very careful where you want to draw the lines to be allowed to proceed, and only replace a layer on the top, low enough that it's not their decision, high enough that most of the mostly white collar workforce is not significantly affected. Miss, and you will be stopped by management, or government.

Just for reference, I encourage to watch this video: The last job on Earth: imagining a fully automated world | Guardian Animations

Answer 2

I'd like to offer an alternative idea: in the scenario you described, where robots become smarter than humans, I suggest that, after a short interruption, nothing changes.

Here's why: in the future you describe, human beings have already designed robots to be smarter than they are. It follows, then, that those robots can now design robots that are even smarter than themselves. If this cycle continues, then in a short while you have robots capable of tasks that human beings cannot even dream of.

So in essence, the robots lift themselves to another dimension of work that is incapable of being rendered by humans anyway, freeing up all of the "menial" jobs that robots previously did (a.k.a, the white-collar jobs). And it just so happens there's a bunch of white-collar workers who were out of a job for a few years, waiting to jump back in.

Answer 3

Will they accept lower-living standards if they take the low-paying jobs that are still available

But they don't have to be lower living standards. Robots are making things more cheaply. Cheaper stuff means increased living standards for a given pay. Now perhaps that won't help people with multi-million dollar salaries, but for someone making close to a break even salary, cheaper stuff means that a moderately reduced salary is actually worth more than the original salary.

You don't define high paying jobs, so let's pick an arbitrary definition: high paying jobs are those that pay more than twice the median salary/wage in the United States. So every job paying more than twice the median is taken by a robot who will cost twice the median per year. For simplicity's sake, all robots are leased and all operating costs are contained in the lease. As a rough estimate, let's call that \$50,000. Roughly 25% of the population makes more than that. Let's assume that 5% of those people are close to \$50,000 and will keep their jobs. The top quintile (20%) in income made an average of \$194,000 in 2014. That's households, so adjust to $100,000 to get individuals.

So 20% of jobs are 50% cheaper. That's a total drop of 10%. So a job up to \$55,555 could be replaced with a \$50,000 job and that person would be just as well off. And everyone who makes less than \$50,000 is strictly better off. And that doesn't even count the ability to increase such workers. If someone paid \$100,000 for one human worker, they might choose to hire two or more \$50,000 robots to do more work.

And I don't believe that captures the real advantage. Currently the highest income people make most of their money from capital. If we replace all of them with comparatively inexpensive robots, we'll reduce the cost of capital. Cheaper capital means higher productivity, more jobs, and cheaper stuff.

Answer 4

I have worked in an engineering company and many people I knew had lots of ideas for really good products, but no time and money to go off and try to make them. Managers are often not interested in pursuing science projects or experimental technology because there is a high risk of loosing money. As a result many engineers produce products way below their potential.

One strategy would be that each household owns a worker robot or two that they maintain that brings in income for that home. The humans in the home would spend most of their time in recreation and creative pursuits rather than work. In that case, those who were engineers may create some really cool stuff now that their time was freed up.

Under that model there would of course be a lot of people who would realize they didn't need to earn a degree to make any money as long as their parents could buy them a couple of robots who would work for them. But hopefully some people would choose to educate themselves anyway.
Societies often make the most advancement when people have the luxury of pursing art and science. One of the reasons Google has been so successful as a company is that they allow their employees to use 10% of their time on experimental projects (one example was google-sketchup).

Alternatively, companies could use the robots like existing companies use entry-level engineers. "go fill out all these forms our process requires". "go review this schematic for errors". "go check that the parts list isn't missing anything". "go design this trivial power supply circuit that we need but I don't want to waste my time on because I am doing the more complex development work". As a result the existing engineers may become more productive because they are freed up from the necessary but less productive parts of their job.