Computers Won't Take Your Jobs, They're Creating New Ones
Thanks to the internet, we've all got an embarrassment of riches in terms of stuff to panic about. Between ISIS, North Korea, our orange-faced president, and the grim nightmare specter of Hillary Clinton's emails, everyone has something to worry over. And one of the most common panic buttons on rotation across the headlines is that ROBOTS ARE GONNA TAKE ALL OF OUR JOBS! JESUS TITBALLS, PANIC!
But is the situation really this dire? I took a deep dive into the evidence and sat down with some experts to try to figure that out. Here's what I learned ...
Historically, Computers Create More Jobs Than They Kill
And that absolutely seems worth panicking about. Unemployment crept above 20 percent during the worst part of the Great Depression. 47 percent unemployment would spell the end of our civilization. If these headlines are correct, we're 20 years away from unimaginable change. But those headlines -- and the study contained in them -- don't tell the whole story.
First off, those researchers are only claiming that 47 percent of all current jobs will be automatable within 20 years. But that automation doesn't happen in a vacuum. For example, the authors of that study make a point about how computer algorithms are increasingly doing the work of lawyers and patent attorneys:
But rather than leading to a vast unemployment crisis for paralegals, the field is expected to grow by 8 percent through 2024. Computers are capable of doing a lot of the work paralegals used to do, which means their employers are able to offer more services, for less, to more customers than ever before. For the foreseeable future, paralegals will get to keep paralegalin'.
"What happens when you automate something? You make it cheaper, you make it better-quality, and that drives up consumer demand." That's Professor Jim Bessen, a lecturer at Boston University who studies the "economics of innovation." He authored a comprehensive study on how automation impacts jobs. He argues that computers tend to increase employment in most fields, and I am almost 75 percent certain he's not an agent of Skynet. Professor Bessen doesn't think much of the 47 percent study: "It's just complete bullshit. They identified jobs like accountant and bank loan officer. There's just too many things that humans do."
I reached out to the authors of that study for commentary, and they never got back to me. (Possibly because Professor Bessen terminated them to aid the rise of the machines, possibly because they thought Cracked was a much more hardcore version of High Times.) But Professor Bessen was able to make some compelling points about how automation's impacted employment over the last hundred-ish years. He pointed out that the impact of machines on the tailoring field wasn't exactly what you'd expect:
"Typically at the end of the 19th century, the average person would have just one set of clothing," but as machines made clothing cheaper to make, "the amount of cloth that was consumed per capita went up 10 times, so that offset the amount of labor required per yard of cloth." It turns out that "for the first 100 years of textile automation, automation was accompanied by growing employment."
You see that in a lot of fields. When ATMs were first introduced, people thought they'd put thousands of bank tellers out of jobs. Here's what actually happened to the employment rate of bank tellers:
It turns out that an entire job field being wiped out by machines is actually a pretty rare occurrence: "I looked at the number of occupations in 1950 and how many disappeared, and of the ones that disappeared, how many could be put down to automation. And it turns out there's only one. The one occupation that can be attributed to largely disappearing because of automation? Elevator operator."
"But Cracked, you machine-fondling hobgoblins," you may ask, "what about self-driving cars? What about all those manufacturing jobs that don't exist anymore? My uncle lost his job to a machine you robo-sympathetic race traitors!" First off, Jesus, chill the hell out. Second ...
The Risk Isn't Unemployment -- It's Inequality
The most common criticism of Professor Bessen's ATM point is that while teller jobs experienced a boom, that boom has started to level off and decline. And Professor Bessen freely pointed out that the boom in textile jobs didn't last: "In 1940s, 1950s, we had almost half a million production workers in the textile industry, and today it's under 20,000. There's been a huge drop in jobs in the textile industry... some of that is due to globalization, but most of it, actually 3/4ths of it, is automated."
Oooooooooooooooh shit, everyone! Get back on the panic bus! The machines ARE going to take all of our jobs. Internet, you can resume freaking out.
Everyone start fucking until we make a John Connor!
Only, maybe don't freak out. The gradual elimination of certain jobs by machine still doesn't mean an unemployment crisis. In 1950, the unemployment rate was 5.3 percent. And our current unemployment rate, after decades of automation ... is 4.3 percent. When you look at, say, the employment rate for manufacturing jobs since 1960, it seems like a disaster:
And for people who loved their manufacturing jobs, it was. But we didn't wind up living in a jobless hellscape where welders were reduced to sucking robotic cocks to pay their rent. Here's Professor Bessen again: "What you find is in manufacturing industries, computer automation still does tend to reduce jobs, but in service, healthcare, and finance, it's associated with growing employment rather than declining."
Employment in the service industry has ballooned from 13 percent to 30 percent in the last 60 years. And that brings us to one of the actual, real worries of automation: Machines aren't going to leave us unemployed, but they might force a lot of us to work shittier, lower-paying gigs. U.S. income inequality is at the highest level it's been since 1928.
"Obviously, the unemployment level is pretty low right now," says Bessen. "It doesn't seem like machines are putting a lot of people out of work, and that's what a deeper dive into the data said. But it doesn't mean everything's rosy and there's nothing to worry about. There are a lot of people who are losing jobs, but more jobs are being created in other occupations and other places, and it's a difficult transition."
The actual quasipocalyptic danger of automation isn't machines taking all the jobs; it's that machines turn middle-income people into lower-income people. They also create a ton of good jobs. Pretty much every article you read on the Internet is written by someone who wouldn't have a job if not for computers. But it's not like someone who's manned a drill press at a Ford plant for 15 years can hop right into being a social media guru or an internet comedy writer. Retraining takes time and money, and someone with kids and a mortgage probably can't afford to go grab another degree when a robot takes their job.
So we do have reasons to be worried about automation. But there's also a reason to be excited, because ...
Computers Are Making Work More Satisfying For A Lot of People
Data gathered from workers in the UK and the U.S. indicates that job satisfaction is on the rise. More people are more content in their careers today. Some of that happiness may be that many folks are just happy to have a job. But it does appear that a lot of people find their jobs more satisfying than they did a decade or two ago.
One reason for this might be that since computers can handle so many rote tasks, the jobs left for people are more creative and thus more fulfilling. This lines up with Professor Bessen's expectations for the future: "You're going to see more creative occupations. We've been seeing that. One of the examples I look at is typesetters and graphic designers. Desktop publishing came along, and it eliminated most of the jobs of typesetters and compositors, but it created many more jobs ... Stuff that would've been done on the typewriter 30 or 40 years ago is much more highly designed. There's much more creative content. You're also seeing growth in jobs where interpersonal skills are important. Independent bookstores have actually been growing, and Amazon itself has actually been growing into having brick-and-mortar bookstores."
Some of the apocalyptic predictions about automation are absolutely true. For one thing, self-driving cars are going to wipe out millions of jobs. Even Professor Bessen expects to see that happen in the next decade or two. It's going to suck for a lot of people, probably a lot of people you know. A good 20 percent of you are reading this in between Ubering people around your city. Computers absolutely give y'all something to fear.
But remember how the service industry blossomed like a cash-filled flower as manufacturing jobs collapsed? Well, we're currently seeing another industry blossom thanks to more and better computers. And yes, I realize the flower analogy didn't add anything to that paragraph. Let's just all move past it and talk about how the entertainment industry is in the middle of exploding. It grew 66 percent from 1998 to 2010.
More of the workforce makes a living off of creating content right now than they have at any other point in human history. Spending on entertainment hit a high point, per family, in 2008. It declined slightly as the recession set in, but over the last few years it's steadily crept back up, and in 2016, we spent more on entertainment than we had in 2008. Digital media has caused a huge surge in entertainment jobs -- and not just in Los Angeles, but as far afield as a whole bunch of African nations.
The fact that hundreds of thousands of people are making money off of YouTube, Patreon, Indiegogo, and Kickstarter doesn't negate the fact that hundreds of thousands more lost their jobs thanks to computers. The Venn diagram of "unemployed smelters" and "YouTube celebrities" probably doesn't overlap much. So the big question we're all going to have to answer in the next 20 years isn't "How do we deal with robot-created unemployment?" It's "How do we retrain people to do all these new jobs before they go bankrupt and wind up eating the rich?"
We're Going To Need To Give Poor People Free Money
This entry is not about universal basic income. A lot of people have written about how universal basic income, a permanent minimum wage for everyone who can't work, will be necessary once the machines take over.
Finland is currently in the middle of a study on how well universal basic income actually works, as are cities in California, India, Italy, Canada, and beyond. It'll be a few years before we have a lot of scientific data on exactly how well it works. In the meantime, the United States also has millions of people who hate paying for public schools. People who oppose a $15 minimum wage probably won't get on board the "free money for everyone" train.
But there's an option in between basic income and letting millions of people go homeless when we finally teach robots how to drive. A charity called GiveDirectly is currently experimenting with what they call "unconditional cash transfers." Basically, they give limited amounts of money to impoverished people -- the equivalent of a year or two's wages -- in the hope that they'll invest this money in their future.
Your inner Republican might assume that giving poor people a sudden pile of cash would primarily benefit the methamphetamine market. But GiveDirectly has actual data on how their cash transfers work. I talked to Johannes Haushofer, an assistant professor at Princeton and an economist for MIT's Poverty Action Lab. He conducted a study on GiveDirectly's cash transfers in Kenya. From 2011 to 2013, they gave randomly selected households roughly two year's worth of expenses. Everything was randomized. In some households, the wife got the money; in other households, the husband. Some households received lump sum payments, and others received monthly payments.
The first thing they noticed is that no one went right out and spent their big pile of cash on alcohol and flat-screen TVs. They spent it on new roofs for their homes, investing in their own businesses, educating themselves and their families, and, of course, buying enough food to not starve to death. The families who received the payments showed long-term benefits: They were less stressed, better educated, healthier, and had more "durable" assets (livestock, a nice home) than they did before the study.
Another cash transfer study, conducted in Bangladesh, showed a 38 percent increase in earnings for extremely poor people who received piles o' money. This study of 10,000 poor people in six countries also found that poor people tend to invest the money they get in cash transfers into their futures, rather than additional six-packs. We don't know yet if universal basic income would work, but there's actually reason to believe that temporary cash transfers might be a better solution to the problems of automation. Haushofer says: "The labor supply response of a basic income grant could be different from a cash transfer. You might imagine that if you just get money once, you know that next year you're on your own again, it doesn't make sense to reduce your labor supply. But when you're good for the rest of your life, you might think differently."
I should also note that Haushofer is still bullish on the idea of basic income: "There are surveys on how people would respond to basic income, and those are generally pretty optimistic in terms of whether people would stop working ... so I don't want to paint it black. I think it's a great idea and should be tried, and I'm optimistic that it'll work well." But the point is that, right now, we know cash transfers can help lift people out of a poverty spiral.
It may be that most Americans will never accept the idea of paying for someone else to get a permanent meal ticket. But you could sell cash transfers as basically paying for people who've lost their jobs to learn how to do new ones. The payment is a stopgap, so they can keep contributing to the economy and not die in the streets while they learn how to go from, say, mining coal to coding. Professor Bessen agreed: "The problem with a permanent basic income may well be that it discourages people from making the effort to retrain. So temporary support seems to be much more suited to the real problem" -- that is, the problem caused by computers.
If we're willing to be smart about it, automation doesn't need to be a humanitarian disaster. And there's at least one way in which it might make the world an objectively better place ...
Automation Might Benefit The Ladies Most Of All
Without getting too political, we can all agree that the last six months haven't been exactly the most inspiring days in the long battle for women's equality. There's a reason The Handmaid's Tale is seen as so relevant right now, and it's got nothing to do with the fact that bonnets are making a comeback.
Barring the establishment of an anti-women theocratic murder state, women are actually poised to benefit the most from automation. This Atlantic video with Jerry Kaplan of Stanford and Saadia Zahidi of the World Economic Forum makes the case that "automation could place a premium on the type of work that women tend to be good at, like person-to-person interaction, reading human emotion, collaboration, and creativity."
When I brought this up to Professor Bessen, he cautioned that "the evidence is weak," but "it also seems quite plausible". He pointed out two findings from his own research:
"Young women were often key in adopting new technologies. For example, it was mainly young women who worked with the new textile technologies in the Industrial Revolution, and also when these industries industrialized in Japan and China. Second, that pattern appears to be true today for information technology; women are more likely to use computers at work, all else equal."
But wait, there's more! In his research on cash transfers, Professor Haushofer also found a surprising benefit for women. The cash transfers lead to a drop in domestic violence, both in the households that received the money and in their neighbors. We don't exactly know why yet. My theory is that toxic radiation bestowed sentience on a pile of cash, which now fights spousal abuse as the Green Knight. Professor Haushofer noted another possibility:
"It could be that, in the treatment households, the husband stops beating his wife for whatever reason, either because she feels empowered now or because he's less stressed and less aggressive. The wife might then share this with her friends, it could echo around the village and become a new norm that it's not OK for the husband to beat his wife. That's one possible channel, but we don't have good evidence yet as to whether that's true or not."
Other research into cash transfers around the world has made similar observations. And this study from the EU noted that unemployment among men led to a decrease in domestic violence, while unemployment among women caused it to increase. There's compelling evidence to suggest that when people have more money, they're less vulnerable to abusive assholes. This, uh, probably isn't that surprising to anyone reading this from Ferguson or South Baltimore. But it is one more reason to look to the coming robocalypse with excitement, rather than dread.
The Age of Machines is upon us, and it might be pretty sweet.
Robert Evans and his videographer, Magenta, have traveled to Iraq three times to film this documentary about the media war against ISIS. It's a story of how memes and social media have changed the nature of warfare, and a story about the brave young Iraqis who risk their lives to wage this war. You can back this documentary right now, on Indiegogo.
Be sure to check out our new Cracked Podcast miniseries, "Looking The Part," in which Soren Bowie and Daniel O'Brien are dissecting pop culture's greatest beards, scars, and tattoos. Listen to part one HERE, and catch parts two and three in the Cracked Podcast feed on Apple Podcasts, or wherever you get your podcasts.
Subscribe to our YouTube channel, and check out How Robots Already Tried (And Failed) To Take Our Jobs, and watch other videos you won't see on the site!
Also follow us on Facebook -- humans only, please.