There'll be flying boats, condos with moats Cultivated oceans, flying cities in the sky Hangin' underneath our bubble No more toil and trouble Singin' bout that sweet ol' by and by - Nitty Gritty Dirt Band, "Turn of the Century"
This is a sentence that has never been said before: you can really miss out by skipping the appendix of an academic paper. This delightful chart is an excerpt from work in Carl Frey and Michael Osborne’s paper “The Future of Employment: How Susceptible are Jobs to Computerisation?”, which predicted the likelihood that a particular profession will be automated in the coming decades. A 1 equals a 100% chance of automation.
Likelihood of Automation (low to high)
|0.13||Urban and Regional Planners|
|0.27||Captains, Mates, and Pilots of Water Vessels||The authors disappointingly fail to predict the likelihood of robo-pirates.|
|0.4||Judges, Magistrate Judges, and Magistrates||I’m amused by the image of human lawyers pleading their case before robot judges. “OBJECTION DOES NOT COMPUTE.”|
|0.43||Economists||Interesting… economists predicting an almost even shot at their own demise.|
|0.6||Slaughterers and Meat Packers|
|0.66||Pest Control Workers|
|0.68||Boilermakers||Sorry Purdue. You’ll have to rework your freaky mascot into an even freakier cyborg.|
|0.89||Bakers||That handles Butchers and Bakers – presumably Candlestick Makers have already been automated.|
|0.91||Tour Guides and Escorts|
|0.94||Accountants and Auditors|
|0.99||Telemarketers||Few tears will be shed. Until you realize that robo-telemarketers are actually even worse.|
This is a very partial list: there are hundreds more in the paper.
These numbers might bring to mind a dystopian wasteland of unemployment and human misery (is your profession near the bottom of the list?).
Frey and Osborne’s work adds to a mountain of evidence that laborers and unskilled workers will continue to lose out in the next decades. Even if only jobs with a score of .90 or greater are eliminated by robot workers, millions will lose out on their jobs – excluded in the table above are dozens and dozens of semi-skilled professions in the 80s and 90s, meaning they are much more likely to be automated than not. Service people in some sectors will become increasingly marginal as automated systems get more sophisticated.
Although common sense would seem to dictate that human workers will face struggles as a result of automation, macroeconomists battle over what the overall impact on society will be. To be sure, workers who can be cheaply replaced with a robot will not get laid off one day and waltz into a new position the next. The “techno-optimists” point out that technological innovation has historically always delivered more employment in the long term. For example, the loss in assembly line jobs over the course of the 20th century was offset by an increase in management, technical, and IT jobs that paid more and were safer.
In one of the great marketing coups of modern academia, adherents of this viewpoint describe their opponents as believers in the “Luddite Fallacy” that productivity gains necessarily eliminate jobs. No one wants to be a Luddite.
Worriers, Neo-Luddites, and many economists argue that this time is different. Unlike the looms of yesteryear that threatened only textile jobs, computer automation affects a huge swath of human endeavors. The Luddites accuse techno-optimists of making huge leaps of logic in their certainty that more jobs will be created over the long term.
Techno-optimists are right to point out that new forms of work have always trended towards less back-breaking labor. I am glad my children will not be weavers or coal miners, much less subsistence farmers.
It’s not very satisfying, however, to rest your argument on the idea that we cannot predict what future jobs will look like and that it will all balance out in the long term. How long is the long term, and will the benefits to society accrue evenly among individuals? Cities with a legacy of manufacturing industries may have fewer dangerous, back-breaking jobs, but technology and software jobs haven’t rushed to fill the void. These cities became so dependent on industry that they have fewer jobs, period. With the lower-skilled workers that have already been replaced by machines, you see declining workforce participation, not a smooth transition to the tech economy.
The types of jobs that are resistant to automation are ones that require high interpersonal or creative skills. Assuming people who currently hold to-be-automated jobs do not have those skills, their best possible outcome is low paid, unstable work. It’s plausible that as work becomes more automated the remaining human jobs in a field will follow the trajectory of “adjunctification” familiar in higher education, becoming unpredictable, with low salaries and no hope for advancement.
Assuming the techno-optimists are right, what will “new jobs” even mean in the context of a more service-oriented economy? Will the bottom 90% have to become masseuses or manservants (this is real) to the upper classes? Button pushers? Bullshit job-holders? Our conception of what work is “highly skilled” is transitory, based on Frey and Osborne’s predictions: today’s accountants and doctors, stereotypically stable and skilled professions, may be easily replaced in certain settings.
When is it Good to Have Less Work?
If wringing your hands over the inevitable march towards human irrelevance and poverty isn’t your idea of fun – weirdo – you may be wondering if there’s anything to look forward to in our robotic future. For most of recent history, machines were supposed to save us labor in ways that would make our lives far more comfortable. Arguably, Ludditism has always been the fringe movement – the easy living of technology-enabled work has more often been a source of hope than despair. Keynes famously said that by 2030 we would all be working a few hours a week with the rest of our time spent on leisure – in short, that our “economic problem [would be] solved”. In the Jetsons, the most mainstream view of the future possible, George finds anything more than his one hour a week shift pushing a button to be hard labor. So if machines really are replacing so much of the work we do, why aren’t we all getting closer to the Jetson workweek?
The standard theoretical explanation is that we found “more things to spend money on” making the “choice” to work fewer hours less palatable. This may be partly true – iPhones appear to have become a base level of Maslow’s Hierarchy of Needs – but this response has always bothered me, relying as it does on a conception of labor freedom that has never really existed. Maybe it looked likely from the 1930s, but we are as far away from it now as we have ever been. The problem is far more systemic than the “we wanted more things” theory gives credit for.
First, there are factors that make people reluctant or unable to have the “labor mobility” predicted by Keynes. First and foremost among these is tying health insurance to working a certain number of hours at your job. This is essentially a historical accident: during World War II wages were capped, so companies started to use health insurance as a sweetener to attract the best workers. After the war, a series of federal rule changes solidified the connection between work and health insurance, and we’ve been stuck with it ever since. The Affordable Care Act’s Employer Mandate is the most recent reinforcement of that forced marriage. The system makes people less likely to leave their jobs or cut back hours if they want more leisure time. It also creates enormous incentives for companies to hire part-time workers without benefits or only allowing full-timers to work 40 hours a week (if I pay their health insurance, I want to get the maximum amount of value).
An even more fundamental problem is that no worker can really “decide” their work hours in our current system. Imagine an interview where you told a prospective employer, “I’m extremely passionate about CorpTech’s mission. I do want to let you know that I’ll only be working 15 hours a week, which will meet my family’s needs and gives me time to work on my hobbies, spend time with my family, and volunteer.” I would love to see an estimate of the percentage of positions that offer meaningful flexibility in number of hours worked and would be shocked if it was greater than 5%. If you’re working ten hours a week in this country, it’s usually not by choice and too unpredictable or low paying to allow you to meet your needs.
The US needs to take a long hard look at what we need to change about the system to make sure people can earn enough to survive and have schedules that will allow them to have meaningful lives. These questions are fascinating to me, and in an upcoming series of posts I’ll continue to look at various policy and philosophical approaches to establishing a better balance between working and living.
If Frey and Osborne are at all correct, it adds a bit of urgency to the calculation, because soon there will be less work and more leisure to go around. In this way, computerization could be both a crisis and an opportunity for rethinking the system. It seems obvious that the incremental reforms – bringing up the minimum wage indefinitely, providing health care through employers – won’t get at this fundamental question.