The Omnivore has written about this before--but now that FiveThirtyEight brings it up again, it's worth looking at one more time. What's that? YOU ARE GOING TO BE REPLACED BY A ROBOT! Nate Silver's new news-engine brings up the plight of the low-wage (fast food) worker who, after a minimum wage increase, is suddenly eclipsed by a robot.
“[T]here was a type of employee at the beginning of the Industrial Revolution whose job and livelihood largely vanished in the early twentieth century. This was the horse. The population of working horses actually peaked in England long after the Industrial Revolution, in 1901, when 3.25 million were at work. Though they had been replaced by rail for long-distance haulage and by steam engines for driving machinery, they still plowed fields, hauled wagons and carriages short distances, pulled boats on the canals, toiled in the pits, and carried armies into battle. But the arrival of the internal combustion engine in the late nineteenth century rapidly displaced these workers, so that by 1924 there were fewer than two million. There was always a wage at which all these horses could have remained employed. But that wage was so low that it did not pay for their feed.” (A Farewell to Alms, page 286)
Would You Like Fries With That, Humanoid?
For those jobs in which at least 25 percent of workers earn an hourly wage below $10.10, the median Frey-Osborne probability is 82 percent [chance of being replaced by a machine]. When we weight each occupation by its total employment — for example, there are a lot more cashiers than sewers, and they have different likelihoods of computerization — the median probability is even higher at 91 percent.The Numbers
The FiveThirtyEight article references a study done by Carl Frey and Michael Osborne that assigns a chance of replacement (in a 20 year window) to hundreds of job categories. For example, Statistical Analysis: .66 or Bus Drivers: .67. How'd they do this? They took a series of dimensions like Originality, Persuasion, Manual Dexterity, Caring For Others, and Working in a Cramped Workspace and assigned chances of technology catching up to humans in their time-frame.
The Omnivore isn't really qualified to gainsay their method (they were explicitly not making discrete assumptions about what technologies will exist in 20 years--something that seems absolutely necessary to determine which jobs are truly replaceable) but if anything it may be under-counting the number of replaceable jobs. Their model predicts that computerization is principally confined to low-skill, low-wage occupations but one or two major breakthroughs could turn that over.
Right now, when The Omnivore flies into New York through LaGuardia Airport there is a vast sea of seats with affixed iPads and Credit Card readers. You can sit down and surf--or order a meal / drinks (Through the iPad) and swipe to pay for them. A human being brings the food to your table--but the entire concourse operates with like 2 or three people.
If they had drones? It might be one or two fewer (they still need a bar-tender probably). Some of the restaurants have seaters (hostesses) but they all have the iPads for the transactions. It's a little cool, yes--but it's efficient and relatively trouble-free.
FiveThirtyEight ties the chances of a replacement to minimum wage--but I don't think that's the driving factor in whether a job gets replaced I don't think that at all.
What Determines If Your Job Gets Automated?
The Omnivore doesn't have a lot of stats or studies--but does have (some) experience with voice and handwriting recognition software and how customer-contact channels work. Here are guidelines that The Omnivore thinks will be important in the coming 2 decades.
- Do customers WANT the human interaction? For some jobs automation becomes a DIY tool. For example, home-care. Imagine that you have trouble bathing yourself: having a human helper for this probably rates on the "requires caring" axis--but what if you had an automated washing-assistant. You wheel up and slide in and then the machine helps you wash yourself? In this case, today, you WANT help--but probably do NOT want a human helper if you don't need one. In the case of a greeter at a restaurant, human interaction is probably a positive. In the case of a server it probably isn't.
- Error/Edge-cases: Voice response is a TERRIBLE fit for a help-line for the basic reason that people generally only call the help-line when things have gone wrong. Having a waiter take my order is not an improvement over an iPad when what I want is a #5 and a coke. When I want a salad--hold the black-olives--but add extra tomatoes ... and can I have just oil and vinegar? It's not listed as a dressing ... THEN I need a person. Even if error/edge cases are only 2% of your business if you get rid of people you're going to have problems. Automation is terrible for dealing with these cases, barring a break-through.
- Improvement Of Service. In the case of an ATM there is an improvement over interactions with a human teller for basic transactions: you can race through the process from many locations without having to get over to the (comparatively very few) branches. If automation has some improvement associated with it then the job is likely to be automated.
I also want to talk about cost. The quote above has horses phased out because paying to keep them is too expensive for the work they do. Note that humans sort of keep and house themselves--horses do not. The expense of keeping a horse, for example, involves bathing it, allowing it room to exercise, and so on. This is, together, probably much more expensive than just "its feed."
Similarly, the cost for automating, say, the back-end of a McDonalds should not just be looked at in terms of wages--but also: you need someone to mop the floor or unclog a toilet? You want to have a manager to take care of an angry customer? You want to tell some kids it's time to leave the parking lot? Automation for making fries and handling the drive-through can't do any of those things--and a single person may not be able to do everything at once. Whether or not you get automated may have less to do with skill / wage and more to do with how many different things a single worker is expected to do.
The Future State
On the other hand, if we were to, say, hit the 80+% mark on that study (meaning that every job for which they have .80 or above gets replaced by machines)? We'd have ... massive unemployment--and it'd be forever / structural (i.e. not because of a downturn or the economy). What happens then?
The future state looks like one of the following:
- Everyone's a genius! Could job training and education give us back those jobs? Probably not. While it's fine to say we need a lot of worker-retraining it's clear that many jobs (like Taxi driver, truck driver, and so on) are the kinds of things lots of people can do--while engineering is the kind of thing not-everyone can do.
- Massive Eternal Unemployment and Starvation: The problem with the "We need a culture of work" attitude is that if you don't have jobs available to people a 'culture of work' alone is not enough to provide actual work. The Left-wing response to this is government-jobs (rebuilding infrastructure). The Right-wing approach to this is to lower barriers to creating businesses so that more jobs are created. Neither of these will work if automation can do the jobs that are created. If we have limited social-net and insufficient jobs there will be starvation. That's 1+1=2.
- Massive Social Services Net: The alternative to starvation is to have social services that provide for people. We see disability doing that today. We also need ... universal healthcare. We probably need scale back some of what we provide to tax-payers as well (the employed) to make everything balance. This is unpopular for a variety of cultural reasons. If you are a Republican, you don't need this explained. If you are a liberal, imagine that I have just cut back police protection and am now charging you a $100/mo "Internet Usage Fee" so that you can read this blog.
It's also worth noting that the problems above don't account for some singularity-grade advancement in either AI or robotics or both (The Omnivore thinks remote quantum computer installations beaming "wi-fi" expert-systems intelligence to robots will provide a platform for more aggressive job replacement than the study probably accounts for).
The Root of the Problem
Fundamentally, the problem with the above scenarios is that the short/mid-term driving market environment is that American society doesn't need producers--it needs consumers. This is the driving reality behind all the economic behavior we see today. What am I talking about? Look:
What these show is that you can have unemployment go up and have both productivity and the stock-market climb as well: short/mid-term, at least, you don't need workers to get stuff done or have profitable companies! In short fewer jobs just means bigger profits without the loss of productivity. Everyone does more with less and it pays off.
Mostly? Thanks to technology.
This suggests that (a) companies will not simply create jobs if given more money--we can see Apple stockpiling more money than some fairly wealthy nations and it still isn't creating jobs in America at a phenomenal rate.
So what's the solution? Well, if you don't like the safety net (if you want to see the future of that, look at the Vancouver soccer riots) The Omnivore has a suggestion for you: Make Work.
The Make-Work Economy
Imagine our future 24-year-old in the year 2030. She wakes up in the morning in Class-S (for 'slacker') public housing and rides the electric tram to work at the Uncle Sam Coffee Shop. The USCS is a Postal-Service like entity that exists to ensure that rain, sleet, or snow American citizens can get their caffeine. Her job? Her job is to take the cup of steaming java from the machine and bring it 22 feet to the customer seating area.
The customers all pay electronically: Using their phones--the USCS CoffeeNOW! App places their order when they click to request it and then geo-locate near the shop. Their order is prepared by a Barista 3000: a fully automated coffee robot that sits in the back room. The coffee costs 10 dollars a cup (BitCoin not accepted--this is a Government run agency!)--but many customers pay most of that with coupons they get for watching TV live and tweeting about it during the show (so that advertisers have some confidence their shows are being watched!).
During her shift, she is monitored by the CoffeeBoss x21, an electronic intelligence agent hooked into the security cameras. It ensures that (a) she does not sit down (posture recognition), (b) that she greets customers and serves them with a smile(!) (facial recognition with attitude reading software), and (c) doesn't surf the net or otherwise goof off during her shift (her government sponsored phone-plan shuts down save for emergency calls--which may be monitored, you know, for quality reasons).
Sometimes the shop needs to be cleaned: our worker can wipe tables down--but the real cleaning is done by a Hy-Genie Automated Cleaning robot that comes every 6 hours to sterilize the environment, pick up trash for recycling, and change the air-filters (since all fossil fuel cars have been phased out and most people use public transit, SMOG alerts have gone way down--but still: hypo-alergenic environments are healthy!).
The USCS has a proud tradition dating back to 2021 when the last Starbucks was nationalized by the Booker Administration under the No Employee Left Behind Jobs Act. It had no employees until President Paul shut down delivery drones for good (hardest hit Amazon--next hardest: the Target Delivery Fleet). The NELB JA required every outlet to have at least one (1) human employee creating literally thousands of jobs for
long term permanently unemployed.
Liberals like this because "it gives everyone at least simulated meaningful work" and conservatives like it because it keeps people off the street and, maybe, "instills a good work ethic."
In any case, the nuclear-grade catastrophe of paying people for "not doing anything" is narrowly averted.
The American dream shambles on ... in radioactive zombie form.