Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
Robopocalypse, Not Yet?
You can believe there’s a Lieutenant Commander Data in our future without also believing he’ll be visiting soon. Economist Robin Hanson agrees with the former speculation, not so much the latter. Hanson thinks “super-robots are likely to arrive eventually” and will “eventually get good enough to take pretty much all jobs.”
Eventually, eventually. But what about right now or pretty soon? What about IBM’s Jeopardy champ Watson, Baxter the flexibly programmable robot, and the Google driverless car? And what about that scary Oxford paper that predicts 47% of US jobs are just a decade or two from being automated away?
Well, there is evidence that automation is already having a big impact on workers, particularly those in middle-skill jobs composed of “routine, codifiable tasks,” according to economist David Autor. And this may be contributing to the “jobless” recoveries of the past three recessions. What’s more, you can thank automation for this simple chart looking at manufacturing employment and output:
And perhaps we are on our way toward a future where a small, tech-adept slice of the population has high-paying jobs while the rest will be physical therapists and high-end butlers — if most have jobs at all. Note the current recovery where GDP is expanding, jobs are being created, but median wages are going nowhere.
But let’s not get ahead of ourselves. Hanson thinks Rise of the Robots: Technology and the Threat of a Jobless Future by Martin Ford does just that, which is why he doesn’t much like it. Ford frets that a mostly jobless, automated world is fast approaching, Hanson writes, and thus we need to soon tax the rich heavily to fund a basic income for the rest of us. The book comes out in May, but, according to the publisher description, Ford argues that “artificial intelligence is already well on its way to making “good jobs” obsolete … The result could well be massive unemployment and inequality as well as the implosion of the consumer economy itself.”
Hanson offers a numbers criticisms: For starters, that 47% job-loss figure is not rigorously calculated. And while median wages have been stagnant and the labor share of income falling, many factors are probably at play. Indeed, Autor recently wrote that “the deceleration of the U.S. labor market after 2000, and further after 2007, is more closely associated” with bursting bubbles and the rise of Chinese manufacturing than computerization. More from Hanson:
But while computer prices have been falling dramatically for 70 years, the job-displacement rate has held pretty steady. This suggests that jobs vary greatly in the computing power required to displace them and that jobs are spread out rather evenly along this parameter. We have no particular reason to think that, contrary to prior experience, a big clump of displaceable jobs lies near ahead.
And then there is Ford’s fourth reason: all the impressive computing demos he has seen lately. … . Only rarely does Ford air any suspicions that such promoters exaggerate the rate of change or the breadth of the impact their new systems will have. … And of course several generations have seen A.I. demos with just as impressive advances over previous systems.
To be fair, I have not read Ford’s book, only Hanson’s critique. But even if Ford is wrong and over the long run technology and mass employment can coexist, “the lessons of the Industrial Revolution suggest that the transition could last quite a while and could be very painful,” as AEI’s Michael Strain has written.
So why not get to work on the smart policies that would equip workers for a more automated future and which are obviously good ideas on their own? Before we crank up taxes and start writing big checks to the forever jobless, how about some of these ideas from Erik Brynjolfsson and Andrew in “The Second Machine Age“: (a) improve education with higher teacher salaries, more accountability, and new digital models; (b) create more startups through less regulation and more high-skill immigration; (c) loosen intellectual property regimes; (d) more government support for scientists, including via prizes; and (e) upgrade infrastructure.
Longer-term, the MIT economists would prefer a negative income tax over a basic income since the wage subsidy “encourages people to start working and keep finding more work to do even if the wages they receive for work are low.” Another option, suggested by economist Tyler Cowen, are so-called universal 401(k) plans where government would help fund tax-free retirement accounts for lower-income Americans. I would also recommend taking a look at these ideas from venture capitalist Marc Andreessen on creating a more dynamic economy.
The prospect of an Age of Automation, whenever its arrival, is a good reason for policy action. Let’s just make sure they are the right policies.
Published in Economics
For an interesting, thought provoking take on this topic, I recommend “Who Owns the Future” by Jaron Lanier.
From my own limited perspective, I witness daily how automation depresses wages of many while increasing the wages of a few.
Jobs that, 20 years ago required some knowledge and skills, today are push-button simple. They no longer require the old skill set and knowledge base. Because the jobs have been ‘de-skilled’, the wages paid to those who do them are commensurate with the lack of skill necessary. Which is not to say that the employees are uneducated. They are not. Most are college graduates. But the JOB doesn’t require their talents. The SKILL now resides in dense lines of computer code. The job is pushing buttons. Sign on here. Upload this. Run these two macros. Download that. Etc.
I hear people at work joke that if their computer only knew the various required passwords and had a finger to type the passwords in…it could do their job without them at all. Then they are shocked that they didn’t get a raise.
The guys/gals writing all that code and the business-line experts advising them are getting paid very well. But the users … Not so much at all.
I’m not sure what set of policies addresses this. But it is a genuine problem and getting worse.
It seems to me that this utopian productivity would create an economy without money before it created an economy without workers. Money would simply be an outmoded method of keeping track of a person’s contribution, and wouldn’t be worth anything, because nothing would be scarce (at least nothing that pertained to a human’s needs).
The catastrophes predicted seem to assume that we will hold onto a scarcity-based economic model in the post-scarcity future. This is not a reasonable situation.
Star Trek had it right. In the future there will be no money, at least not in the sense that we use it or understand it. We need money to keep track of scarce resources to meet our needs and to allocate those resources better. When those resources are not scarce there won’t be a need to keep track, and allocating money (as a holder of value) based on their production or use won’t make sense.
However, these lofty predictions don’t deal with the very real transitional pain. To that point, robots can’t produce creativity. So it seems that the way to protect yourself during the transition is to be creative, since the robots have you beat in the productivity department, but can’t hold a candle to human creativity, and won’t ever be able to.
Hi JordanW.
Good point about creatitivity. Do you think the gains in ‘creatives’ can make up for losses eslewhere?
And are we making a mistake with our current emphasis on STEM education? Perhaps we should be emphasizing art, design, music, film/video etc. along with entrepreneurship?
Said it before, and I’ll say it again.
Automation is a diminishing-returns proposition. The easiest, cheapest stuff to automate away is always the worst jobs.
People are competing with robots. If you want to improve people’s competitive edge, cut out the government. Wonks always want to implement policies. You’re better off excising the policies you have right now.
If you lose your job to automation, get a different job. If you don’t have the skills, then get the skills you need. Nothing about this is new with computers.
I don’t think so, but in and of itself STEM isn’t enough. You can’t replace the programmers due to the software-complexity problem, software is getting ever more complex, and requiring more software to solve the problems which other software creates. So its this kind of weird recursive problem where the solution requires more solutions, which require more solutions, etc.
I think the emphasis on SCIENCE! is overrated currently. Scientific careers suck right now.
Engineering is very important, since having ideas is one thing, but bringing them into reality is where it’s at. This is not unlike the software programmer in many respects. Making the thing happen is just too important.
Math is probably one of the areas where humans and computers interact in the kind of ideal way which I imagine them to work later. The humans do the high level thinking, and the computer does all the brute force computation.
In general, I think you need the ability to work with computers at some level, but more in the sense that you make the computer work for you, not that you work for the computer.
There’s two problems with this assessment. One is the assumption we will never duplicate human intelligence, and the other is, even assuming robots can’t technically be creative the way we think of humans, that they can’t convincingly fake it.
This sort of analysis, that automation will lead to systemic unemployment, flies in the face of the totality of economic history which has been more output, more and better jobs and more employment.
Yes, I assume that human intelligence cannot be made in machine form. I don’t think this is an unreasonable assumption though. At the very least, getting machine-produced creativity will occur significantly later than humans with robots can achieve post-scarcity production levels.
I imagine what the future looks like is a human’s mind in control with a machine’s accuracy, precision, and speed.
It’s a genuine non-problem.
It’s the defining characteristic of capitalism, and of all human innovation since the dawn of the hand ax.
Everyone’s boogeyman: the ATM
Evil computers
I gave the example earlier of CAD. CAD is the ultimate “job destroying” technology. What required hundreds of engineers and dozens of support staff…now can be done by 1 guy on 1 computer.
Results? HUGE increases in the demand for engineers. Weird.
So, what are the “policy implications” here?
Get more skills and become more knowledgeable.
Same “policy implications” as have existed in humanity since around 30,000 BC.
PS: Of course, the funny thing about the arguments so-called “conservatives” make against automaton and machinery (and virtually all so-called “middle class” arguments)…
…are purely Marxist arguments.
Returns accumulate to capital and not to labor :)
Of course, this is ignoring the fact that there is such a thing as…human capital…and that their predictions don’t correspond to what we have observed for the past 200 years of human history (or 30,000 years of human history)
But why let such a small thing as observed reality stop us?
“Its a genuine non- problem” unless you are one of those whose working a job that has been de-skilled.
Anecdotal evidence….
Just this morning in a meeting with a potential client. Topic: upgrading their current system to handle increasing volume and some technical tweaks. Just before wrapping up the client said…
“Since you’d already be working with the system, think you could tighten up our _______ processing? We’d like to codify and automate lots of what the processing team does.”
This particular team has several dozen ‘processors’. These are good, solid middle class jobs. Full time, decent pay, benefits, 401k, vacation, small bonus at Christmas. Nobody’s getting rich, but nobody’s on food stamps either. I started my career in one just like it. You have to know a little spreadsheet software, a bit of database stuff, be detail oriented, know a bit about the business, the products, the clients and what they do. You need some skills and some base of knowledge.
After we were done discussing what, exactly, they wanted this ‘codifying’ of the process to entail, it was clear that pretty much anyone off the street could do the new, automated job. We can encode exactly what happens in scenarios A-V. Pretty straightforward to do. This makes up 85% of their business. No more skilled/knowledgable processing staff required for these. Just people to verify that the system is doing it’s thing. Scenarios W-Z and anything else oddball will get routed to a skilled staffer.
There will end up being almost as many jobs, but clearly the number of ‘skilled’ staff required will be drastically reduced. The rest will get downgraded a coupl’a three, four pay grades. Maybe not right away. But certainly as people leave, the replacements will be these low pay grade new hires.
And people will continue scratching their heads about why median income isn’t rising given the low unemployment rate.
Imagine how many more jobs they can create if they got more efficient systems ;)
By this logic, imagine how many more jobs we would have if we still used hand axes to build things, instead of machines! Everyone would have “good middle class job” :)
Imagine all the jobs their customers will be able to create as the products they make become cheaper as a result, increasing demand in the market.
But this is the history of humanity: automation.
And yet this is precisely the process through which the “middle class” was created.
How can this same process “destroy the middle class”???
What those “median wages aren’t rising” arguments miss 2 very important things:
1) What can you buy with those wages? Meaning, if 50 years ago the median wage was higher (let’s assume), that doesn’t tell us anything as to how much they had to work to be able to afford the same thing as today.
If things are getting cheaper (and they most certainly are), on a comparative bases (i.e. the same good vs the same good), then even if wages remain flat, what you can buy with them increases.
2) Where these jobs are being created. The same job in NYC will require 3 times the salary as the same job in North Carolina. Yet the South is the one where the jobs are growing. But obviously, employers can afford to pay a lot less in these areas, because the cost of living is so much lower.
I.e., the same job, the same person…will get a lot more out of a paycheck 1/2 the size in North Carolina as they will get from a paycheck 2x more in NYC.
Hence, adjust for cost of living…what have wages done?
I agree that Jaron Lanier’s books on this are insightful and well worth reading.
The issue with the “Star Trek Thesis” is that in Star Trek they didn’t just have automation, they also had something called a “replicator”, which was a magical device that created food, clothing and other necessaries from apparently nothing in particular. That kind of device eliminates need to a large degree, but I don’t think anyone is predicting the creation of a credible replicator type device for things like food and energy, and certainly not within the timelines that robot-influenced automation will hit. Star Trek also didn’t have AI, apart from Data (who was not in the original concept), which is another issue, but a more distant one than robotic automation.
The robotic automation wave won’t eliminate money. The bots will still need energy to run, and there are still substantial limits on the supply of energy for that. The makers of the bots, the programmers of the bots, and the providers of energy that the bots use will make a killing, and they will be paid in cash. In addition, as I point out above, people will need cash for food, clothes, housing. Where will it come from when their jobs are automated? At first, it will be the low hanging fruit jobs that will be automated, but it won’t stop there. People with the required skills will of course thrive (ability to design and program the bots, ability to proficiently maintain them, etc.), but that’s a limited number of people. The transition to something new will be undoubtedly rough.
30,000 years of human history says otherwise. After all, that’s all we’ve doing for 30,000 years: eliminating jobs through technology.
All that’s done, is increase everyone’s lot in life, substantially.