Japanese Engineers pulled off an impressive demonstration earlier this year. They arranged a convoy of four freight-hauling trucks, cruising smoothly at roughly 50 m.p.h., maintaining a steady distance between the vehicles as they roared past commercial buildings and rolling fields. The impressive part was that three of the trucks had no driver. Only the lead vehicle had a human at the wheel; the others were controlled by lasers, software and wireless communications. The automatons were really good too. At one point, the engineers arranged to have a chase car accelerate past two units of the convoy and cut in behind the second truck. The third and fourth vehicles promptly fell back to a safe following distance–then closed the gap after the chase car left the line.
Watching a video of that demonstration, I couldn’t help wondering how many people are employed as truck drivers in the U.S. According to the Bureau of Labor Statistics, the number was 1.6 million in 2010 and is projected to grow to nearly 2 million by 2020. These aren’t bad jobs, either. Median pay for truckers as of 2010 was about $38,000, with many veteran drivers earning twice that, plus benefits. Were those Japanese trucks a vision of doom bearing down on all those middle-class jobs? In today’s world (let alone tomorrow’s), the speed and force of technological change seem to summon an uh-oh for every gee whiz.
The automation of human labor is as old as the Industrial Revolution. From the steam engine and the cotton gin to the desktop computer and the robotic welder, machines have enabled leaps of efficiency that create far more jobs than they destroy. And yet many economists and technologists believe that things are different this time, that society is entering a new and troubling phase as computing power and other advances make possible the creation of ever-more-powerful robots. What if the economic growth of the future produces more jobs for more robots, leaving humans behind? What if we’re heading toward a future in which a handful of creative humans marshal an army of ever-more-intelligent machines while everyone else languishes? How does the world work without … work?
We can see jobs vanishing before our eyes. Airport ticket counters used to teem with employees; now only a handful stand by to assist customers who use kiosks. Travel agencies once brightened streets with posters touting sunny destinations; now people book vacations online. For workers from accountants to X-ray technicians, technology has chipped away at a virtual alphabet of occupations. This revolution is likely partly to blame for the slow growth in jobs since the crash of 2008. Consider the manufacturing sector: robust growth in production has produced little or no growth in employment. According to a study published by the Federal Reserve Bank of Chicago, by the end of last year U.S. manufacturing had recovered 75% of the output lost in the Great Recession but only about 21% of the lost jobs.
Whether you lean toward the gee whiz or the uh-oh, there is no doubt that the workplace is heading into uncharted terrain. As MIT professor Erik Brynjolfsson and principal research scientist Andrew McAfee put it in their book Race Against the Machine, “Computers (hardware, software, and networks) are only going to get more powerful and capable in the future, and have an ever-bigger impact on jobs, skills, and the economy. The root of our problems is not that we’re in a Great Recession, or a Great Stagnation, but rather that we are in the early throes of a Great Restructuring.”
The Human Advantage
Truck drivers are part of a huge web of workers around the planet who sort, pack, load, transport, unload and deliver the astonishing bounty of the global economy. Every touchscreen, sewer pipe, negligee and air compressor, every tractor, picture frame and Xanax tablet. Shipping is the lifeblood of human commerce, employing tens of millions of workers to pump goods and materials through the world’s transportation arteries. If robots can steer trucks down highways, they can certainly pack boxes and load pallets. An enormous number of jobs are at stake in the shipping and distribution fields, which makes them a good coal mine in which to hunt for canaries.
As it happens, I don’t live far from the Overland Park, Kans., headquarters of YRC Worldwide, a trucking giant that operates 450 shipping terminals, 15,000 trucks and 52,000 freight trailers across North America and beyond. The company’s clients come from every economic echelon, from Walmart to Wally’s Widgets. CEO James Welch showed me to a chair one recent morning in his modest 10th-floor office. His eyes gleamed when I mentioned a claim by those Japanese engineers that robotic convoys could cut fuel consumption by 15%. That’s a lot of money in a highly competitive business–in the neighborhood of $100 million per year for YRC, which spent $630 million on fuel in 2012. Robots would also cut payroll costs, and machines don’t take sick days or file for workers’ compensation. Yet Welch was underwhelmed by the prospect of self-driving vehicles.
“I wouldn’t bet against it,” he ventured between sips of a soft drink he picked up at a nearby gas station. “But it doesn’t seem very real to me anytime soon.” Why? Because unlike the trucks in Japan, which cruised along empty roads marked with freshly painted white lines to guide the laser-driven steering, YRC’s trucks traverse the wild and often crumbling roads of the U.S.’s cities and countryside. “I get around enough to see that we have some pretty serious infrastructure issues in this country,” Welch continued, “not to mention all the people who are driving while texting or eating or watching movies.” How’s a computer to know that the driver of the car merging from the on-ramp is busy posting a Vine video of herself blowing bubble gum? How does an algorithm deduce that the cars up ahead are swerving to avoid an unseen tire in the middle of the road?
“I could picture it working if we built separate roads for driverless trucks,” Welch said, “but who is going to pay for that? If they pass the cost to the industry, the tolls would be cost-prohibitive. And even if you had dedicated lanes, what happens when it’s time to pull off for fuel?”
Welch told me that YRC teaches its drivers-in-training that truckers must make an average of 200 observations and 20 decisions per mile, and while those numbers are mere child’s play for a computer, the tricky part is how many of the decisions are unpredictable. They rely on experience and intuition. This is why, for the foreseeable future, YRC’s ideal truck driver isn’t a robot. It’s Arthur Cage of Memphis. In nearly 40 years of hauling and delivering freight, Cage, 64, has covered 5.6 million miles without an accident of any kind. The secret to his success starts with his attitude, he told me when I called him at home: like a computer, he never loses his cool or lets his concentration flag. But unlike a computer, Cage “always [expects] the unexpected.” He studies the weather for hints of a slick road or a gusting wind. He peers into the cars ahead and beside and behind him, trying to intuit the next lurch or swerve or stomping of brakes before it happens. He’s always learning. Sometimes he will follow an erratic driver to the next rest stop so he can add to his understanding of unpredictable hazards.
“I asked a man once, ‘Back at mile marker 39, when your car started wobbling all over, what were you doing?’ And he said, ‘Lighting a cigarette.’ You see? He and I were both lucky that the shoulder was firm where he ran onto it. If that man had done that same thing where the shoulder was made of gravel, taking his eyes off the road and weaving and bobbing, he would have lost control of that car. My job is to anticipate that.” When Welch ponders the future of YRC’s workforce, he thinks less about robots than about hiring and training young versions of Cage.
Yet there are many varieties of work under the big umbrella of the U.S. shipping industry, a nearly $17 trillion field, and many of them are robot-ready. Highly standardized and repetitive tasks are ideal for automation. Take loading and unloading cargo ships, for example. The vast majority of oceangoing freight travels in standardized intermodal containers, and robots are tirelessly adept at stacking and unstacking these boxes, moving them from ship to freight train to flatbed truck. Engineers envisioning the ports of the future foresee a continuing decline in the number of longshoremen, whose ranks have been hard hit by automation. At the Port of New York, for example, their numbers have been cut by some 90% since peaking in the 1960s.
Similarly, small-package handlers like Amazon and FedEx deploy robots to perform many of the tasks at their distribution centers around the world, because both firms work with standard-size boxes and envelopes. Last year Amazon paid $775 million to acquire Kiva Systems, a company that makes robots, zippy little orange mules that race through warehouses storing and retrieving merchandise.
YRC exists in a different world. Every shipment presents a new piece of an ever changing puzzle: how to pack all sorts of stuff, from long steel rods and bundles of lumber to barbecue grills and lightbulbs, neatly into outbound trucks. “It’s always something different, and we haul this freight through traffic, along city streets, and eventually back our trucks right up to our customers’ docks,” said CEO Welch. “We’ll be hiring drivers and forklift operators for as long as I can see.”
Replaced by Robots
A distinction made by Frank Levy of MIT and Richard Murnane of Harvard, a pair of economists who have studied the impact of automation on human employment, helps explain the next phase of the Great Restructuring. If your job involves learning a set of logical rules or a statistical model that you apply to task after task–whether you are grilling a hamburger or issuing a boarding pass or completing a tax return–you are ripe for replacement by a robot. As computing power continues to increase and code writers continue to work their binary miracles, more and more occupations will fall into that category. We will see more robots in teaching, sales, nursing, stock trading–and journalism too.
The surviving jobs will be of three kinds, Levy and Murnane predict: solving unstructured problems, working with new information and carrying out nonroutine manual tasks. It’s hard to design a suitable robot to style hair, fix plumbing problems, design buildings or plot corporate strategy. A computer might be able to write a routine contract, but attorneys who can write persuasive legal briefs will be in demand for the foreseeable future.
I wondered how the same question might look to someone whose business is robots. So I reached out to the editor in chief of the Journal of Field Robotics, Sanjiv Singh, a longtime faculty member at the cutting-edge Robotics Institute of Carnegie Mellon University. Field robotics means exactly what it sounds like: the study of real-world applications for the theoretical investigations of other scientists and engineers. In his workshop, Singh has tackled a wide range of applications. He spent one patch of his career applying robotics to harvesting fruit. Currently, he’s perfecting large-scale autonomous air vehicles, which he helpfully translated as “giant helicopters that can fly by themselves.”
Interestingly, Singh endorses a number of Welch’s ideas. “It takes a long time to develop these technologies,” he says, and the idea of robots as artificial humans is not something he expects to see. Instead, he thinks of humans and robots primarily as collaborators. Robotic devices can relieve people of jobs that are “dull, dangerous, dirty,” whether those jobs are on mind-numbing assembly lines or involve tasks in unpleasant environments, like clearing land mines or welding on the ocean floor. (A major robotics competition this year, sponsored by the Department of Defense, will produce a humanoid robot designed to climb ladders, patch leaks and shut off valves at disaster sites like the Fukushima Daiichi nuclear power plant in Japan.)
In other cases, technology can enhance people’s ability to perform tasks. Rather than imagine a day when truck drivers are obsolete, Singh prefers to think about ways that sensors designed for robotic applications can support safer driving by humans. “I recently bought a Volvo with an optional drowsy-driver detector,” he says. “It beeps at you if you start to veer. That’s not what laypeople think of when you talk about robotics, but it is from our perspective.”
Can sensors and software help fighter pilots land safely on storm-splashed aircraft carriers? Can machines speed apple harvests and relieve migrant workers of their bushel baskets? Can devices in the walls at Grandma’s house ensure that she takes her medicine, turn off the stove when she forgets to and summon help if she breaks a hip? Yes, yes and yes. What worries many economists, however, is the next step, where robots go from assisting human workers to making them obsolete.
That step, says Singh, will be fraught with difficulty in many fields. Here again, driving is a good example. For years, Google has been spending huge sums (for reasons not entirely clear) to develop self-driving cars. And it has made great progress, racking up more than 300,000 test miles without an accident, according to the company’s official blog. Yet the nation’s cab drivers need not worry about their jobs. The Google cars may be self-driving, but they require humans at the ready to take over in case of trouble. The cars also fare poorly on snowy streets and can be flummoxed by such routine hazards as cones in the road. “There’s still a long road ahead,” Google acknowledged last year.
Nor are these issues purely technical. The closer we get to self-driving vehicles or pilotless airplanes or human-free restaurant kitchens, the more questions of law and culture kick in. “Like all machines, some are going to fail, and how are we going to handle that? Who’s held responsible when a driverless car kills someone?” Singh asks. And what about emotional acceptance? Self-driving trucks hauling loads of ore back and forth at a remote mining operation in Australia are one thing. (Yes, it’s happening.) “It’s something else to put your daughter in a self-driving car on a rainy night and send her off to her violin lesson,” Singh says.
The Permanently Unemployed
“Creative destruction” is the phrase coined by economist Joseph Schumpeter to describe the irresistible force by which market economies grind weak, fading or inefficient industries into fertilizer for new, stronger ventures. In 1870 it took nearly half of all working Americans to run the nation’s farms and feed its people, according to Census data. Today far more food is grown by a tiny fraction of that number: less than 2% of U.S. jobs are in agriculture. Yet the destruction has led to many more jobs than were lost. The lightbulb helped kill the New England whaling industry, but the workers employed thanks to electricity could fill enough ships to darken the seas.
As ever-more-sophisticated machines move into all manner of workplaces–both blue collar and white collar–entire categories of work may disappear. “Workers will have to make a historically unprecedented transition,” Martin Ford, a Silicon Valley software designer and writer, explained in a recent article for the Association for Computing Machinery. “Rather than simply acquiring new skills and moving to another routine job, workers will have to instead migrate to an occupation that is genuinely nonroutine and therefore protected from automation.” Mastering creative, analytical or intuitive work is not for everyone, and yet it may become the only path to a productive occupation. Those who cannot make the leap may be doomed to jobs that pay too little to be worth automating–cleaning hotel rooms, trimming shrubbery–or face permanent unemployment.
Even the job of creating robots is not a safe 21st century career. As you read this, there are some very smart scientists in a discipline called evolutionary robotics who are busy developing robots that invent other robots. The limitations of the driverless truck give us reason to think their task may take a long time to complete. But the fact that they are working on it at all shows how deep we are into the robotics revolution. While we don’t know what effect it will have on our economy or how we interact with one another, it should spur us to work just as hard to prepare ourselves, our children and our communities to meet the challenges this revolution will bring.
More Must-Reads from TIME
- Inside Elon Musk’s War on Washington
- Why Do More Young Adults Have Cancer?
- Colman Domingo Leads With Radical Love
- 11 New Books to Read in February
- How to Get Better at Doing Things Alone
- Cecily Strong on Goober the Clown
- Column: The Rise of America’s Broligarchy
- Introducing the 2025 Closers
Contact us at letters@time.com