This article will appear in the Spring 2017 issue of The American Prospect magazine. Subscribe here.
Five to ten years from now, Uber hopes, the following will be an everyday occurrence: A driverless Uber car pulls up at Nick and Nicole Smith’s house at 7:30 in the morning. Their two kids, Julia, 16, and Joey, 14, hop into the car, and it then drops them off at school.
The Uber car returns to the Smiths’ suburban home, picks up Nick and Nicole, and drops Nicole at the train station to catch the 7:55 into the city. The car then drives 40 minutes to drop off Nick at his company’s headquarters nestled in an office park. During the drive, Nick uses his laptop to answer emails and finish a PowerPoint.
At 5:00 that afternoon, a driverless Uber car picks up Julia from field hockey practice and Joey from baseball practice and takes them home. Meanwhile, Uber has sent another car to pick up Dad at the office park at 5:30, and that car makes it to the train station in time to pick up Nicole, who’s due on the 6:15. The car then stops at a Chinese restaurant so the Smiths can pick up a takeout dinner. It waits while Nick runs in to get the food, and it then drops off the Smiths at home.
Welcome to the brave new world of driverless cars and trucks. In this not-too-distant future, many Americans are likely to conclude that they no longer need—or want—to own a car. No more worrying about tune-ups or getting gas or shopping for auto insurance or renewing registrations. Uber’s controversial co-founder and CEO, Travis Kalanick, is very high on driverless technology, saying, “When there’s no other dude [i.e a driver] in the car, the cost of taking an Uber anywhere becomes cheaper than owning a vehicle.”
And Uber is far from the only player. Ford Motor and BMW say they’ll have driverless cars on the road within five years. General Motors invested $500 million in Lyft (obtaining a 10 percent stake) to help it develop driverless cars. Tesla, Volvo, Daimler, Toyota, Fiat Chrysler, Google, and Apple have also joined the race. And now mighty Intel, eager to get into a new, fast-growing sector, jumped into the fray in March by announcing a $15.3 billion deal to acquire Mobileye, an Israeli company that is a leader in making sensors and cameras for self-driving vehicles. Depending on how things shake out, Ford or Fiat Chrysler may own and manage self-driving fleets that compete head to head with Uber.
Otto-mated: One of Otto's self-driving, big-rig trucks leaves the garage for a test drive during a demonstration at the Otto headquarters on August 18, 2016, in San Francisco.
The coming of driverless cars and trucks raises several big questions. How long will the transition to these vehicles take? What will be the benefits and the costs to society? Who will be the winners and who the losers (and it seems clear that many of the five million people who work as drivers will be losers)? And what role should the government play in managing and regulating the arrival of this new technology, or should all this be left to the free market to sort out?
Also, what role should government take to help the millions of truck drivers, taxi drivers, and others who might lose their jobs, or should these displaced workers be left to fend for themselves? Should society require the winners from this new technology—most likely some giant corporations and the consuming public—to somehow share their gains with the losers, perhaps through some special levy on these new vehicles to finance programs to help the losers?
The shift to driverless cars and trucks will be one of the most important new technologies since the horseless carriage replaced buggies. Yet the advent of self-driving vehicles will be different from the arrival of many previous technologies. Because we can see this technology coming years ahead of its arrival, we as a nation are in a position to take some smart policy steps to prepare and help the many people who will lose their jobs as a result of this new technology. But given America’s dismal history with easing economic transitions, such smart steps may not be taken, despite years of advance warning.
Self-driving vehicles will threaten the jobs of as many as five million people—workers who make a living as taxi drivers, long-haul truckers, Uber and Lyft drivers, local delivery drivers, limo chauffeurs, and even many bus drivers. Beyond that, driverless cars could mean an end to many gas stations, auto repair shops, car washes, auto dealers, commercial parking lots, and car rental agencies, not to mention traditional auto insurance. Moreover, because this new technology will lead to a more efficient use of cars, the nation could well need fewer autos—and that could reduce overall car production.
These cars will benefit the public in ways that many people don’t yet appreciate. “Most of us in the U.S. commute all alone in a single car—the average American driver spends 51 minutes to and from work every day,” says Raj Rajkumar, a professor of electrical and computer engineering at Carnegie Mellon University. “This is wasted time and wasted productivity. Driverless cars would be a big win for productivity. You can spend 45 minutes doing something productive, including taking a nap.”
Rajkumar also sees a sharp drop in auto accidents. “About 1.3 million people die every year worldwide from auto accidents,” he says. “In the U.S., it’s about 38,000. Most of these accidents are caused by human error. Most of these lives would be saved with driverless technology.”
Rajkumar adds that autonomous cars would be a godsend for blind, disabled, and elderly Americans who can’t drive. “These people would gain mobility,” he says. “It would greatly help the 1.5 million legally blind Americans and the more than five million disabled people who can’t drive.” Autonomous cars will make it far easier and, in theory, cheaper for many disabled people to get to work. And they will make it easier for many elderly and disabled people to go to movies or parks or shopping, and for the elderly to see their grandchildren and care for them after school.
LAWRENCE KATZ, A LABOR ECONOMIST at Harvard, says there’s no denying that this radical new technology will imperil many jobs. The five million drivers who could lose their jobs, he says, represent almost 3 percent of the workforce and don’t, as a group, have bright prospects if they get laid off. “It’s disproportionately male, disproportionately non-college,” he says. Five million jobs, he explains, “is about the same size as the decline in manufacturing jobs since 2000. It’s a comparable potential shock.”
The move to self-driving cars will of course produce some new jobs—in engineering, software, and marketing. But displaced former drivers generally won’t qualify for these white-collar jobs. Instead, they could find themselves with lower-paying jobs at Walmart or Target, or they could be trained for something better.
According to Katz, there are two big reasons that the economic pain from losing five million driving jobs will be less than from losing five million factory jobs. First, drivers’ jobs are spread fairly evenly across the United States, while manufacturing jobs were concentrated in one devastated region, the Midwest. The more-even geographic distribution should make it easier for laid-off drivers to find jobs. Second, Katz noted, because factory workers average around $20 an hour, while drivers average $13, laid-off drivers would take less of a hit than factory workers if forced to take $10-an-hour jobs at Walmart. Still, Katz added, “This will be a shock to the labor market for non-college men. They’re already in tough shape.”
Bhairavi Desai, executive director of the New York Taxi Workers Alliance, which has 19,000 taxi, Uber, and Lyft drivers as members, shares these concerns. “Our people see this as a threat, and we know it’s closer than we realize,” she says. “These are workers on the margins, with language issues and low wages. Drivers are an older workforce. It’s people who are mainly in their forties and fifties and older. It’s going to be really hard to relocate these people within the economy.”
“It’s going to be very disruptive,” says Erik Brynjolfsson, director of the MIT Initiative on the Digital Economy. “It’s going to happen faster than a lot of people realize.”
How fast will it happen? Some companies, like Uber and Ford, see driverless cars becoming popular within five years, while some skeptical experts say it will take 20 or 30 years not just to develop a truly safe self-driving technology, but also to get needed government permissions, and to overcome public anxiety that could morph into public resistance.
Uber's Travis Kalanick speaking in London.
Moody’s Investors Service predicts that self-driving technology will not become a common option in conventional vehicles until 2030, and not become standard in all vehicles until 2035. Moody’s further predicts that not until 2045 will a majority of vehicles be driverless and that by 2055, they will become nearly universal.
Uber and Google have prototypes on the road now. It’s one thing, though, to have a few dozen or hundred closely supervised experimental vehicles on the road, but a very different thing to have several million self-driving cars on streets and highways across America. Karl Brauer, executive publisher at Kelley Blue Book, an auto research firm, predicted that such cars would become mainstream across America in four to ten years. But Joan Claybrook, former head of the National Highway Traffic Safety Administration, says it could take 10, 20, even 30 years of tinkering with and improving self-driving technology until it is truly safer than human drivers.
In terms of its impact on the job market, “a huge amount depends upon the rate of change,” says David Autor, an MIT economics professor. “If this is gradual, it’s not a big deal,” he adds, explaining that if this happens over 15 or 20 years, many drivers should have time to retrain and the economy should be able to absorb them without major problems. “But if this happens overnight, it’s a very big deal.”
FOR THE MOMENT, THIS technology still faces sizable hurdles. It doesn’t work when snow covers lane markers, and there’s considerable worry about what happens if the internet system controlling these cars suddenly crashes. Government approvals could be delayed for years if there are more fatalities like that of the Tesla test driver who died in May 2016 when a truck cut in front of his self-driving vehicle. To help make sure that driverless vehicles are safe, huge infrastructure spending will be needed, such as installing road signs that communicate with these vehicles and upgrading GPS systems so they are accurate to within a few inches to help self-driving cars stay in their lanes.
The huge outlays needed to finance these infrastructure improvements raise concerns about cost and about fairness. Supporters of driverless cars argue that such cars will help consumers—affluent and non-affluent—because they will in theory be cheaper than using today’s taxis. In some ways, however, the public investment needed to make a nationwide system of driverless cars possible will be a subsidy for affluent families like the Smiths who will be able to afford to use them far more than poorer families. The billions that government spends on infrastructure to make driverless vehicles safe and reliable could crowd out spending for mass transit, hurting less-affluent families who rely on subways, trains, and buses.
Many technology experts say major infrastructure improvements will be needed to ensure that driverless cars are safe. The electronic equipment—sensors, cameras, radar, lasers—that keeps self-driving cars from hitting other cars relies on lane markings to operate, but when there is fog or when snow covers lanes or when there is blinding glare from the sun, those sensors often can’t function. To remedy this problem, there is a push to develop a far more accurate GPS system that can locate vehicles to within four inches—helping keep them in lanes. This upgraded GPS will likely need an expensive, elaborate, and frequently updated mapping of the terrain.
As a further step to ensure safety and guide these cars, industry experts are calling for a new generation of road signs and traffic lights that contain elaborate electronics to communicate with driverless cars—perhaps telling them exactly where the next exit is or warning of construction ahead or an upcoming lane merger. These infrastructure improvements will be expensive, and one can be sure that the industries rushing to introduce driverless cars will want the government to pay. In the past, the government has paid for GPS satellites and highway signs.
Another infrastructure requirement: The millions of autonomous cars that will eventually be on the road will need to communicate with the cloud, and that will mean significantly expanding the capacities of the cloud. This cost is likely to be borne by companies like Uber, Google, and Ford, and not by taxpayers.
Self-driving vehicles will threaten the jobs of as many as five million people—workers who make a living as taxi drivers, long-haul truckers, Uber and Lyft drivers, local delivery drivers, limo chauffeurs, and even many bus drivers.
Brauer of Kelley Blue Book has other, nagging fears, among them that computer hackers might sabotage a car’s computer controls. For example, hackers might direct 20 driverless cars on a bridge to suddenly turn right, and they then disastrously plunge off the bridge. Or there might be an unexpected technical glitch that causes five cars to blow through an intersection and take out a group of pedestrians. Such a serious glitch or hacking incident could traumatize the public and cause regulators to delay for years giving needed government permissions.
Brauer sees another obstacle to rolling out this new technology. “A lot of people won’t buy into it on a philosophy and on a trust level,” he says. That creates another challenge—many traditionalists may shun driverless cars and keep on driving their own cars. That could mean big problems meshing the supposedly safer driverless technology with cars driven by often-flawed human beings, some of whom will inevitably drive drunk or carelessly change lanes and crash into a self-driving car in their blind spot.
There is also the big issue of how driverless cars should be regulated. Autonomous cars present an auto-safety issue far more momentous than the battles over airbags or cameras for when one drives in reverse. But in the face of industry eagerness and rapidly changing technology, the Obama administration punted. Without much public debate, the administration gave a fairly bright green light to driverless cars in September 2016, when the Department of Transportation issued guidelines permitting self-driving cars. These guidelines call for driverless cars to meet 15 safety standards and called for the 50 states to develop uniform regulations for these cars. There was no notice of rulemaking, and no formal set of rules.
Under these guidelines, Uber, Google, Ford, and other companies could move ahead in introducing driverless cars onto the roads, without any further federal authorization needed—although the government emphasized that it would look over the companies’ shoulder and intervene if it saw safety problems with this radical new technology. In the event of safety problems arising, the government reserved the right to step in and set standards for some technologies, like electronic sensors, that autonomous cars rely on.
Perhaps predictably, some auto industry executives protested that these guidelines were too strict because they raised the possibility that the government would delay or block some new self-driving technologies if it saw safety problems. Currently, the auto industry self-regulates in introducing new technologies, like automatic braking, under the belief that automakers—not wanting to hurt their image or sales—wouldn’t add a new technology unless they were convinced it was safe. But in light of the industry’s checkered history on safety, some skeptics are calling for greater regulation of self-driving vehicles.
In Claybrook’s view, the Obama administration was far too hasty. “There will be some good actors and some bad actors,” she says. “The question is whether the public can rely on anything.”
States have asserted some regulatory authority, but that is largely being preempted by Washington. The states will continue to be responsible for inspecting cars and licensing drivers and insurance. “The way it’s going to work,” Claybrook says, “if the National Highway Traffic Safety Administration does nothing more, the industry can do whatever it wants on driverless cars. The federal government will have opted out.” And of course the Trump administration is even less likely to intervene.
TRUCKING PRESENTS A paradoxical variation on the theme. Last summer, Michael Belzer, a professor at Wayne State University and one of the nation’s leading academic authorities on the trucking industry, was predicting that self-driving trucks wouldn’t arrive for 20 years. He said in an interview back then, “The technology is just not there.”
But now Belzer says self-driving trucks might arrive in ten years, perhaps sooner. This past October, Uber’s Otto subsidiary caused Belzer to shorten his timeframe when an Otto self-driving truck equipped with $30,000 in software and hardware hauled 50,000 cans of Budweiser from Fort Collins to Colorado Springs. With a human driver sitting in the cab as a backup, the truck drove 120 miles on I-25, although the human did get the truck onto the interstate and took over again before the truck exited back onto city streets.
Some experts predict that self-driving trucks will become widespread before self-driving cars do, because trucking companies are so eager to cut costs and have these expensive vehicles run nearly around the clock. “The trucking industry has a huge interest in not having to pay for drivers,” Claybrook tells me. For some long-distance trucking trips, companies may cut back from two drivers in the cab to one, and there’s talk of having convoys with two or three trucks, with a driver in the front truck, and one or two autonomous trucks following right behind.
“One of the big questions should be what’s going to happen to this workforce,” Belzer adds. “A lot of these drivers voted for Trump. What’s going to happen to them? Our tradition in the United States is to hell with these workers and dump them. We get rid of that work, get rid of those jobs, and you, the worker, figure out what to do.”
For now at least, officials with the International Brotherhood of Teamsters aren’t voicing much alarm about the possibility of truck drivers losing jobs. The reason: Teamster officials assert that self-driving trucks will continue to need a driver to help out in case of emergencies, to handle paperwork, to assure the safety of the cargo, to pull a truck into a warehouse. An analogy is aircraft. Planes have had autopilots for generations. Today, aircraft can virtually fly themselves—and do so in bad weather. But despite futuristic projections of pilotless planes, no serious person is pushing to have commercial jetliners without human crews capable of taking the controls.
Kara Deniz, a Teamsters spokesperson, says, “We don’t see anything indicating that these vehicles will be able to function in real-life conditions, go long distances, especially through a city environment, without a human being. You hear these doomsday scenarios about jobs, but there are so many questions that still need to be addressed, like not knowing the reliability and safety, and then there are hacking issues.”
Sam Loesche, a Teamsters lobbyist, adds, “We have not seen a situation where any responsible operator would agree that there’s a role for this technology without an operator in the cab. Our view is [that] the best piece of safety technology is the driver.”
While insisting that drivers will continue to be needed, Loesche switches gears and says that in any conversation about self-driving vehicles, “workers’ jobs need to be a top issue.” He adds, “You can’t have the number of job losses you did in other industries and not see a role for policymakers and the government.”
The main industry group, the American Trucking Associations, shares some of the Teamsters’ views. Pointing to issues like hazardous cargo, driving in cities, and interacting with warehouses and stores receiving merchandise, Chris Spear, the association’s president, told a federal transportation panel in January that ATA believes drivers “will retain an important role in trucking for the foreseeable future, even with automated trucks, just as pilots are still present in airline cockpits.”
Google Co-Founder Sergey Brin shows then-Secretary of State John Kerry the sensor displays inside one of Google's self-driving cars in Palo Alto, California.
Kalanick, Uber’s CEO, has a surprisingly optimistic—though somewhat self-contradicting—take on what driverless technology will mean for jobs. He says that in years to come, Uber will offer a mix of rides, some provided by drivers, some by self-driving cars. Kalanick foresees skyrocketing demand for improved transportation and says this would mean a continued need for drivers—for instance, to help many disabled or elderly passengers in and out of vehicles.
“Many people predicted that the ATM would spell doom for bank tellers,” Kalanick says. “In fact, ATMs cut the cost of running a local bank so more branches opened, employing more people.” Kalanick adds, “Self-driving Ubers, for example, will be on the road 24 hours a day, which means they’ll need a lot more human maintenance than cars today.”
NUMEROUS CORPORATIONS stand to make a fortune from this new technology, and many consumers will benefit, too. “Well-funded companies see a lot of potential revenues at the end of this rainbow, and they all want to get there first,” says Brauer of Kelley Blue Book. “There’s lots of money to be made producing the cars, controlling them, managing their data systems, and collecting information about passengers. Companies will know when he’s going, where he’s going, what he’s doing. Talk about opportunities for targeted advertising.”
Once autonomous cars become popular, Brauer adds, “there will be an explosion of new business models with billions of dollars up for grabs.” With driverless cars, Uber and Lyft won’t have hard-working drivers keeping 60 percent to 80 percent of the fares. Instead, Uber and Lyft will keep the lion’s share of fares, although some might go to taxes and fees and some might go to Google or whichever company runs the driverless cars’ data systems. “Their business model would go from having huge profit potential to having massive profit potential,” Brauer says. “The fight over whether the drivers are independent contractors or employees and the related litigation, all that goes away. It becomes an extremely desirable business model.”
In their book, Driverless: Intelligent Cars and the Road Ahead, Hod Lipson and Melba Kurman point to another problematic effect of driverless cars—companies and their shareholders will grow richer as drivers grow poorer. “As machines take away human jobs,” they write, “one potentially devastating effect will be to further exacerbate the trend toward inequality, a growing global problem.”
Pit Stop: Uber testing driverless cars on Pittsburgh streets, 2016.
MIT’s Brynjolfsson argues that we as a nation should make sure we find a way to share this prosperity and progress with the displaced drivers. The most important step, he said, will be to provide first-class retraining for them. Many economists say bigger and better retraining programs will be the key, even though the nation’s retraining programs have a less-than-stellar history because they often train unemployed workers for fields where there aren’t enough job openings.
Help for laid-off drivers can take many other forms, perhaps including something akin to the trade adjustment assistance received by factory workers who lose their jobs to imports. A beefed-up version would be desirable, because trade adjustment assistance has often been underfunded and has not done enough to support and retrain laid-off workers. Harvard’s Lawrence Katz offers several other suggestions: wage subsidies, an increased earned income tax credit, and, if need be, a government-backed public jobs program. (That last idea is anathema to conservatives.)
A big question is where we, in tax-averse America, would find the money to finance programs to assist displaced drivers. One option would be to impose a tax on self-driving vehicles—perhaps a tax on every mile traveled. At some point, Congress could step in and tell the many corporations eager to introduce—and cash in on—this technology that it will agree to a nationwide green light for all 50 states on one condition: that it enacts a law creating a special per-mile tax to finance programs that aid and retrain displaced drivers.
“Even though this may not lead to mass dislocation,” Autor says, “it may further constrict the set of opportunities for people who don’t have specialized skills. Not to say that driving doesn’t require specialized skills, but it’s not up there with engineers or computer programmers.”
Autor sees demand for skilled construction workers, skilled medical technicians, and skilled repairmen as opportunities for displaced drivers. As for those who do not retrain and develop special skills, Autor says, they might have to turn to lower-paying jobs as home health aides, security guards, or food service or cleaning workers.
Brynjolfsson believes laid-off drivers should be retrained for many “different kinds of jobs and industries.” Many could be trained for sales or marketing jobs, he says, adding that there are huge unmet needs, in particular, in health care and other nurturing and caring jobs. But, he acknowledges, “a lot of truck drivers aren’t going to find much of this appealing. It will be a rough transition.”
America’s retraining programs have often fallen short because they failed to place workers in jobs after months or years of retraining—often because the programs ignored whether there were opportunities in the fields for which laid-off workers were being trained. While education—whether obtaining vocational training or associate’s or bachelor’s degrees—would of course help laid-off drivers, it could prove frustrating if there aren’t enough jobs out there.
An ambitious program to address the nation’s urgent infrastructure problems is badly needed, and that could create a million or more construction jobs, as well as long-term jobs maintaining and operating all manner of “smart” infrastructure systems, not to mention the newfangled auto repair stations and parking lots that will be needed for company-owned fleets of driverless cars. The nation’s renewed focus on preserving and creating factory jobs could also provide valuable opportunities for laid-off drivers. And it is of course important to strive for a full-employment, higher-wage economy that offers job opportunities not just for laid-off drivers, but for everyone. This is a long-term planning challenge that goes well beyond the question of how to re-employ the upcoming wave of displaced drivers.
No one should expect a seamless transition for the many drivers whose jobs get run over by driverless technology. That’s where wage subsidies, earned income tax credits, extended unemployment insurance, and some form of adjustment assistance might come into play. Some advocates argue that universal basic income would be an ideal program for drivers thrown out of work and having a hellish time finding another job. But most drivers are eager to work and probably won’t want to just get by on a modest UBI.
The development of driverless cars is a perfect opportunity for some smart public policy. We as a nation have years to plan, and there will be a rich revenue stream to draw from. Let’s hope that we, the world’s richest nation, adopt some smart, foresighted policies so that millions of drivers don’t get banged up in a head-on collision when this radical new technology takes the road.