Job automation roundup
Many articles about technology replacing labor, the future of the human workforce, and the potential for mass unemployment (and what that means for the welfare state) have been making the rounds lately. I'm listing some of the best here (some are older, but still very important) so there's a single place for them - and because it is one of the most important issues in the world, and will only continue to get more important:
Why workers are losing the war against the machines
Krugman and Delong on automation
Structural problems in the US labor market
American workers race against the machine
A post of my own on the subject:
http://www.opir-music.com/blog/policies/will-robots-steal-your-job-the-future-of-the-welfare-state/
The future of work, part 2
Very interesting and important video. It starts getting especially interesting around 1:00:00 when Cowen and Ford start talking. A few things:
- The idea of "personal shoppers" being a growth area seems dubious. With recommendation and predictive desire systems getting better all the time, I think this is probably a no-go.
- Nannies might be a growth area for a few decades, but decent AI will likely obviate the need for these as well. I'd give this one 50-50 odds over the next 4-6 decades.
- Basically everyone on that panel endorsed the guaranteed income. The fact that people from so many different perspectives had their policy prescription converge to that is really striking.
- This also brings us back to the "how do we occupy the unemployed masses?" question yet again. Arts and leisure activities will likely not need much support from society, as they will just happen. Sports might (stadiums, etc.) Other things we'll probably be looking at will be extremely powerful, long-lasting, side-effect-free (or tolerable side effect having) drugs. That road of course leads to wireheading, but I think that denying that as a possibility now is extremely naive.
A cultural shift will be required, as several people in the video make clear. Having one's self worth (and respect of one's peers/society) be defined by a traditional support-one's-livelihood "job" is going to have to disappear. Demonization of those "lazy, ne'er do wells" is just not going to work in a future where much human labor is not needed.
Another thing this brings to mind is the effect on the labor market that does continue to function: a great deal less unemployment, at least for a long while. For those skills that are essentially not able to be automated any time soon (let's stay over, the next 50-100 years) those that stay in the labor market will not be doing much competing for open positions. It may in fact be an employee's market in those cases where someone is not simply their own business.
Will robots steal your job? The future of the “welfare” state
we're going to need to do some more serious big thinking about more than "work", which is too narrow now. We need to figure how how we're going to "occupy" people in the transition from post-industrial/service/information technology society to a roboticized, post-scarcity, arts and leisure society. If handled poorly, "social unrest", mass protests, and outright violence may be become a regular part of the landscape, what with millions of always-idle, impoverished people just sitting on the sidelines, ignored. How long could this last? One hundred years, perhaps? That's a long time to have constant social upheaval.
Sci-fi has dystopias full of rebellious robots, human vs. robot warfare, grappling with what it means to be sentient, etc. but have startlingly little that deals with a much more realistic question: what does a society where human labor is being made redundant look like in terms of day-to-day human behavior?
I've said in this a number of times in the past, but this position is still at the fringes. Over the coming years, the consequences of mass automation and even "stupid" AI are going to come to a fore. Another entrant in the "automated future" analysis genre appeared recently: "Will robots steal your job?" It's definitely worth reading.
One thing I would like to see done sooner than later, is the analysis of what I touched in my quote: the occupation of people in a world where most human labor is unnecessary. Encouraging the arts and music, competitive games, and various leisure activities, are probably a given (and not much encouragement is likely to even be necessary, since many will do these things on their own given the ability to do so.) The other piece of the puzzle is the one that's really going to need a great deal of thought: how do we support them, and just as importantly, how do we run an economy where 40, 50, 60 of the population does not work (and thus, under the current system, have no income)? The most obvious answer is to move towards a massive expansion of what we currently call the welfare state along with planned population policies (hard, but probably necessary.) The big difference? We'll be not just supporting jobless people, but the economy itself (we can call it econofare or something.) What kind of economy can you have with few consumers? Not much of one. It will require a completely different perspective, one where it's not a bunch of unfortunate (or lazy, depending on your perspective) jobless people, but a bunch of people we essentially pay to be consumers. Items which are scarce will need special handling for sure (hopefully things like food and shelter can be made superabundant sooner than later), but for everything else, it'll be all about simply keeping the flow of money going. The main policy prescription here is the tax-free guaranteed income.
Edit: Two types of jobs I would add to the "won't be automated anytime soon" list:
- Jobs that exist to make others feel powerful/superior. Waiters, house-servants, massage therapists. They do have jobs that could certainly be automated, but you don't get the same feeling of "lording over" others with robots that you do with people.
- Interdisciplinary generalists (i.e., modern Jacks-of-all-trades.) This job isn't a single "job" at all, but a collection of jobs requiring the understanding and ability to synthesize knowledge from different (and sometimes disparate) areas. We will continue to automate parts of specific types of cross-discipline "tasks", but we won't be automating the big-picture viewers.
Does everyone need a college degree?
Does everyone need a college degree? is a well-written article on a study of the US education system, and how badly broken it is. It touches on some things which have been floating around a while in the econo- and political blogs. It's worth a read, but I think that it misses a few bigger picture issues, many of which are related to race-to-the-bottom offshoring and, more importantly now, increasing automation:
- Helping people make the connection between higher education and "what they want to do in the future." What people "want to do" may be jobs that do not pay enough to support a living or have disappeared/are on the verge of disappearing. This is made worse by the fact that jobs that seem stable right now may be gone in just a few years. "Making the connection" to something that doesn't/won't exist isn't very useful. It's sometimes hard to predict what's going to be in demand next, but the whole "mess around for years, read the BLS site after they've figured out the next big thing 5 years after it actually starts, then race back to college and incur massive amounts of debt trying to catch up and by the time you're done the industry/job is gone" "system" is just not working.
- Job training for "middle skill" jobs has the same issue as the item above. Are these jobs really safe? Sure, they are hard to offshore, but they too can ultimately be automated, and even if they can't anytime soon, just how many electricians can a country support? This really needs to be thought through.
- Finally, what we really need is to do some more serious big thinking about more than "work", which is too narrow now. We need to figure how how we're going to "occupy" people in the transition from post-industrial/service/information technology society to a roboticized, post-scarcity, arts and leisure society. If handled poorly, "social unrest", mass protests, and outright violence may be become a regular part of the landscape, what with millions of always-idle, impoverished people just sitting on the sidelines, ignored. How long could this last? One hundred years, perhaps? That's a long time to have constant social upheaval.
Of course, this is all from the purely shorter-term economic cost-benefit perspective. A highly educated workforce is extremely valuable for both a properly functioning liberal democracy, and for an innovative society. With a universal, free, distance-learning focused higher education system, this calculus changes a great deal. We should strive for this.
