One of Uber’s self driving cars was recently spotted by SFist while being driven (the old-fashioned way) through the streets of San Francisco.
Over 30 different companies are working on some form of self-driving vehicle, with the highest profile entry coming from Google, which currently boasts over 1.4 million miles of driverless testing since debuting their first car in 2009.
Until recently, self-driving cars have had an impeccable record, with only a handful of minor incidents to report, each one involving human error on the part of other drivers. The streak ended on February 14th when one of Google’s self-driving cars collided with a bus at low speed while changing lanes to avoid sandbags in the road.
For defenders of the status quo this was a sure sign that self-driving cars aren’t yet ready for primetime. Google dubbed it a learning experience, implementing 3,500 new tests to make sure this sort of collision doesn’t happen again. Still, for supporters of the technology, the issue isn’t to eliminate accidents entirely, but rather to establish that self-driving cars are safer than the system we currently have in place.
Following the February collision, BBC News interviewed U.S. Secretary of Transportation Anthony Foxx about the incident. Foxx stated:
“It’s not a surprise that at some point there would be a crash of any technology that’s on the road. But I would challenge one to look at the number of crashes that occurred on the same day that were the result of human behavior. I think the question here isn’t comparing the automated car against perfection, I think it’s a relative comparison to what we have now on the roads which is you and I, and our eyeballs, and our brains.”
Given that self-driving cars are already safer under certain conditions, and an increasing number of U.S. cities are receiving funding for “smart” technologies, it’s only a matter of time before self-driving cars are commonplace on city streets. However, the path to widespread adoption will require overcoming consumer fears, establishing new regulations, and demonstrating the many ways autonomous vehicles can make life easier.
Are drivers willing to give up control?
When it comes to driving, people tend to overestimate their own abilities. Most people agree that texting and driving is unsafe, but yet people continue to do it because they’re convinced they’re the exception. Research has shown that the more confident someone is in their ability to multi-task behind the wheel, the more dangerous they actually are.
Regardless of whether a computer is capable of superior judgment, how do you change the perception of someone who thinks they’re an excellent driver?
A recent poll found that 51 percent of registered U.S. voters would not ride in a self-driving car. The safety of the technology will be irrelevant if the majority of people aren’t willing to cede control of their vehicle in the first place.
The same poll found that only 25 percent of respondents said they would ride in a self-driving car, with the other 24 percent saying they didn’t know or didn’t care. No matter what the actual safety statistics are, the fear of relying on a machine to do something most people feel (overly) confident they can do themselves, could discourage large amounts of the population from using autonomous vehicles.
And that’s not even taking into account people who simply love driving.
Although younger people are increasingly disinterested in cars, there’s still a significant number of people who consider driving—and in some instances their vehicle of choice—an extension of their identity.
Younger people are more willing to give self-driving cars a chance, but even among 18-29 year-olds, only 38 percent say they would buy or lease a self-driving car. For people 65 and over, the number dips to just 12 percent.
In theory, elderly drivers have the most to gain, since diminished motor skills could be supplemented with the assistance of driverless technology. Unfortunately, distrust of technology and fear of change pose a potential roadblock.
Among all ages, there’s a persistent fear of road safety, computer glitches, and how self-driving cars and traditional vehicles can coexist, with over 75 percent of respondents considering each of these items a “concern.”
Finally, even if drivers can be swayed toward self-driving cars with rapid advances in technology and compelling data about increased safety, the auto industry seems to have a vested interest in slowing adoption.
For an industry already combative toward innovation, it’s not hard to imagine manufacturers fighting desperately to keep people driving as long as they possibly can. After all, we want to make sure self-driving technology is 100 percent safe before letting it share the roads with millions of flawless drivers, right?
How do we define the new rules of the road?
Accidents will happen. They already do.
Over 32,000 people die in the U.S. each year from motor vehicle accidents. Even if self-driving cars aren’t perfect, shouldn’t they be measured against the existing imperfect system?
By most estimates, self-driving cars already have the ability to make roads safer. With driver error accounting for 94 percent of all auto accidents, eliminating human fallibility from the equation would greatly reduce the number of accidents. According to a study by the Eno Center for Transportation, if 90 percent of cars on American roads were autonomous, accidents would fall from 5.5 million a year to 1.3 million.
The larger issue is determining who should be held responsible in instances like Google’s recent collision with a bus. Assuming an autonomous vehicle causes an accident, should the owner be held responsible for something that might have been the fault of the car’s programming? And, inversely, should the manufacturer be blamed for accidents that might have been prevented if the vehicle’s driver didn’t intervene prematurely or otherwise supersede the car’s program?
The questions surrounding the insurance and tort liability implications of self-driving vehicles are complex, and there are competing thoughts on how these questions should be addressed.
As attorney Ralph Jacobson wrote, “In the short term, the preferential tool for dealing with liability and compensation issues arising from the use of driverless cars would be one most in tune with, and least disruptive of, the civil justice system’s current operation.” In his article, Jacobson recommended updating the California vehicle code so that self-driving cars are included in the language and fault could be determined much as it is now.
On a state level, these fixes would work. On a national level, there’s lots of work to be done.
Self-driving cars are currently governed by a hodge-podge of varying regulations at the state level. California, Nevada, Michigan, Florida, and Washington, D.C. have adopted laws regulating how the technology is tested and sold, but these laws aren’t consistent across states.
According to Chris Urmson, technical leader of Google’s self-driving car project, legislators in 23 states have introduced 53 bills regulating autonomous cars in recent years, but a big picture approach is needed.
“If every state is left to go its own way without a unified approach,” says Urmson, “operating self-driving cars across state boundaries would be an unworkable situation and one that will significantly hinder safety innovation, interstate commerce, national competitiveness and the eventual deployment of autonomous vehicles.”
As an example, California recently determined self-driving cars will be required to have a licensed driver, while the National Highway Traffic Safety Administration has made no such distinction.
Everyone agrees that ensuring the safety of self-driving cars is important, but defining what that looks like and agreeing on universal standards will be difficult.
Autonomous vehicles also face the momentous challenge of planning for scenarios in which there’s no ideal solution. Likened to the “trolley problem,”—an intellectual exercise in which an observer is forced to choose between allowing an out of control trolley to hit five people or flipping a lever that switches the trolley to a different track, hitting someone who was previously out of harm’s way—self-driving cars will need to be programmed to make tough choices.
In essence, the vehicle’s programmers are in the position of playing God.
Google’s Chris Urmson downplays the trolley problem, saying, “It’s not possible to make a moral judgment of the worth of one individual person verse another — convict versus nun. When we think about the problem, we try to cast it in a frame that we can actually do something with.”
Urmson says the system is engineered to prioritize avoiding vulnerable road users such as pedestrians and cyclists, then other vehicles on the road, and lastly things that don’t move. Unfortunately, this will be of little comfort to someone who is injured by a self-driving vehicle that prioritized someone else’s safety ahead of his or her own. That might account for why three out of four people still have concerns about autonomous vehicles.
It will take time for self-driving cars to win over a largely skeptical public. But, as “smart” technology continues to improve and the benefits become increasingly accessible, the common concerns about self-driving cars should begin to be weighed against our currently flawed system.
What’s next for self-driving cars?
Given the prevalence of rideshare options like Uber and Lyft, it’s not difficult to envision a future in which a driverless car can be summoned with the push of a button to pick you up and deliver you to your destination.
Uber is already testing its first autonomous vehicle in Pittsburgh, with CEO Travis Kalanick pointing out that, “When there is no other dude in the car, the cost of taking an Uber becomes cheaper than owning a vehicle.” Are we heading toward a future of vehicles on-demand?
As The Economist pointed out in their article “If autonomous vehicles rule the world,” despite being one of the most expensive things most people own, cars sit idle approximately 96 percent of the time. For city-dwellers with short commutes, a reduction in rideshare costs could be all it takes to cut the cord on car ownership completely.
The same article also suggests a future in which “some places will probably ban ordinary cars on safety grounds, starting with city centres, resorts, business parks and campuses.”
What about concerns over vehicle safety? Citing a report on self-driving cars from Morgan Stanley, The Economist predicts that attitudes will quickly shift from “I don’t want to share the road with robots” to “I don’t want to share the road with other human drivers.”
In addition to eventually freeing up drivers to tackle other tasks while in transit, a study from the Texas A&M Transportation Institute estimates that a 90 percent penetration rate of self-driving vehicles would double road capacity and reduce delays by as much as 60 percent.
Although a shift from 51 percent saying they would not ride in a self-driving car may take time, it seems silly to bet against technology with that much potential to make everyday life easier. After all, in 1996 only about 6.1 percent of the U.S. owned a cell phone, now that number is around 91 percent.
By most accounts, fully self-driving cars like the current model from Google are still a few years out. However, semi-autonomous cars are already making remarkable headway, with Business Insider predicting “10 million self-driving cars will be on the road by 2020.”
By witnessing the merits of self-driving technology in incremental stages, consumers will be more likely to embrace fully autonomous vehicles when they eventually come to market.
It’s hard not to daydream about the possibility of never having to relinquish your keys because you’re getting too old to drive safely. Similarly, it’s encouraging to imagine a future where one person’s momentary lapse in judgment behind the wheel doesn’t result in tragedy.
The future of self-driving cars isn’t up for debate; it’s simply a matter of how soon.