By now, most tech blogs out there have had the obligatory conniption over White House Treasury Secretary Steve Mnuchin’s comments. Mnuchin told a crowd in Washington, “In terms of artificial intelligence taking over the jobs, I think we’re so far away from that that it’s not even on my radar screen. I think it’s 50 or 100 more years.”
Every tech writer in the US immediately jumped on Mnuchin to ‘harrumph’ and correct him. That’s all well and good. But in the long term, how do we stop this problem from happening in the future? That is, how do we teach people to recognize that technology is, indeed, shirking the job market and will continue to do so?
How Not To Do It…
The first thing we have to do is stop saying “Robots are taking our jobs.” Because when you say that to the average Joe Sixpack – or somebody of even lower intelligence, like a politician – they picture it like this:
…which actually isn’t how it works at all. There is almost never a one-to-one replacement of one human per machine, and even if there is, it would be silly to constraint the machine to the exact same function and limitations as a human.
Instead, technological employment looks more like kudzu. Day by day, inch by inch, it grows and consumes. You don’t notice it happening until one day you step back and notice that your entire garage has disappeared under the landscape. The ways in which automation and A.I. replace jobs are even more subtle than that, because kudzu is just one easily identified plant, while technological unemployment happens in shifting layers and combined fronts. We should be explaining it more like this:
Condition #1: Incremental innovation chokes out jobs gradually.
Case in point: Banking. Consumer banking has seen two major innovations eat jobs. One is the ATM, and the other is the debit card.
ATMs are an easy example. Before ATMs, in order to get cash out, you would have to physically visit the bank during bank hours, stand in line, and withdraw money. Now it’s a matter of sliding a card and punching in a PIN, and money magically comes out of the slot. This reduces the need for bank tellers. When was the last time you went into a bank? It was probably to open the account in the first place, or sometimes to deposit a two-party personal check or attend to some other occasional errand. The older generation of today can remember a time when the line at the bank was this long queue that snaked back and forth – “You’d think you were at Disneyland!” observed comedian George Carlin.
The other innovation is the physical debit card itself. Because before debit cards, your main method of paying a bill was to mail in a check, and your checks would get processed on a huge document processing machine like this:
This gloriously noisy IBM 3890 Document Processor, is an example of what’s commonly called a “transport.” It reads the MICR and OCR on your check and billing statement, automatically reads it into the computer system, then stamps a bunch of processing data on it and deposits it in one of those 32 or so pockets stretching to the back. Those pockets eventually get collected and shipped off to different branch banks, Federal Reserve districts, and so on. The present author can testify to this, because a transport is what I used to operate at the processing headquarters of a major multinational bank – before, of course, getting laid off.
The transport, itself, put many manual accountants out of work when it came along. But when debit cards became popular, it took about exactly one generation for checks to mostly phase out. Along with that, the advent of the World Wide Web and the technique of direct deposit also took bites out of the check processing industry. We consumers used to order checks by the box. Now the book you get from the bank with your welcome bundle when you open an account is almost a lifetime’s supply. Paying by check these days seems quaint, like you should be doing it by candlelight with a quill pen and a wax seal.
Condition #2: An industry can lose jobs to both internal and external innovation.
It’s no secret that brick-and-mortar retail stores are shuttering at an alarming rate. Along with Sears, Macy’s and other mall-bonded department stores, the most surprising is GameStop. Seriously, it’s 2017, and the number one video game retailer in the country can’t keep the lights on? This slow, steady march of irrelevant commercial space is mostly the result of online shopping. It is simply too convenient to click a button and have a product ship itself to your door. This is an example of an external source affecting an industry.
But within retail, we really are seeing some small margin of jobs lost to one-to-one replacement robots. So you’ve probably noticed these self-checkout machines:
…which is an example of internal innovation taking a job. This affects the volume of retail jobs at a slower rate; no matter how good we build these machines, some people just balk at ringing up their own stuff. Even if we went full self-checkout, how does the store guard against thieves, vandals, and other human annoyances?
Nevertheless, together the two effects squeeze retail jobs from both ends at the same time. Fewer people shop at retail stores because they buy more goods online, and even when they go into a brick-and-mortar store, they can use self-checkout, so the store can get by with one or two fewer cashiers.
Condition #3: Shirking jobs in one industry also impacts jobs in many dependent industries.
Let’s talk about those self-driving cars, particularly those trucks that the Wired article uses as an example.
Self-driving cars are trotted out a lot now as the poster child for technological unemployment, and they’re actually not a good example. Incidents like Uber’s accident dim the public’s faith in self-driving cars. Meanwhile, the always-hilarious The Oatmeal webcomic gives us a first-hand experience of riding in a self-driving car – but notes the technology is balky and quirky, outmatched on anything but a smooth test track free of snow and potholes.
Nevertheless, those self-driving cars are coming, and there will be no arguing against them. Traffic accidents in the US claimed 35,092 lives in 2015; US annual auto accident deaths have hovered in the 30K-50K range since the middle of the 20th century. How do you argue against saving 30,000 lives in a year? Never mind whether the self-driving cars will come – after awhile, it might become illegal to drive manually at all.
But in the meantime, driverless cars aren’t working at the consumer level, but one field of driving is ripe for A.I.: interstate trucking. Almost all of interstate truck driving looks like this:
Who needs strong A.I.? You could almost replace the driver with a rope tying the steering wheel in place and a brick on the accelerator. It’s miles and miles of straight highway as flat as a tabletop. Self-driving trucks would definitely be able to handle this, at least in ideal driving conditions.
So we can say that truck drivers will lose jobs to automated truck driving in the near future. But what nobody seems to notice is that we’ll also lose a lot of jobs in dependent industries when that happens:
* Roadside motels and diners – Automatic drivers don’t need to eat or sleep. So there’s laid-off waitresses, cooks, maids, and hotel managers.
* Sleeper cabs – The inside of some truck cabs are furnished like a small R.V. But no drivers means no need for truck cab models with a fridge and cot, so part of the truck design industry gets laid off.
* Traffic cops – One less driver means one less traffic ticket to write, albeit we’re sure cops will find something else to do with their time instead.
* Driver’s Licenses – Automated driving will likely have a standard of controlling software to take the place of individual licensing.
These are all secondary industries that will feel the external blow from the advent of A.I. drivers. As more and more vehicles are driven by automatic pilots, whole swaths of industries will be made obsolete. We can kiss drunk driving laws goodbye, tear down the auto insurance industry, get by with fewer collision-repair shops, and surely close down a few funeral homes when 30,000 fewer people die every year. Sure, that last observation is grim, but hey, a job’s a job.
Not only is technology advancing faster than most people think, it’s even advancing faster than the can think of ways to describe its pace. If we’re going to get political policy-makers to understand this, we’re going to have to try harder to reach out to them.