In recent weeks, BMW and Ford each announced ambitious plans to have fully self-driving cars on the market by 2021. Richard Holman, the head of foresight and trends for General Motors, has said that he expects to see self-driving cars available to the public by 2020.
But consumers still tend to see autonomous cars as a wild idea that fits better in a sci-fi novel than their own garages.
Although self-driving cars receive substantially more media coverage now than they did two years ago, says Brandon Schoettle, project manager in human factors at the University of Michigan Transportation Research Institute (UMTRI), the public isn’t any less apprehensive about them.
In UMTRI’s April 2016 survey of 618 licensed U.S. drivers, 37% were “very concerned” about the prospect of riding in a fully self-driving car, about the same (36%) as in 2014.
A little under half—45.8%—preferred no self-driving, and 38.7% preferred partial autonomy—a self-driving car where the driver can still take the wheel. Only 15.5% preferred fully autonomous vehicles like the Google car, which don’t have brakes or a steering wheel for the driver to take over.
An Alix Partners survey in June 2016, of 1,517 adult drivers in the U.S., asked the questions differently—framing the new technology more as a wish-fulfillment than a worry—and the responses were considerably more optimistic. 73% said they would like a self-driving car to "handle all their driving needs."
“In the last few months there’s been a lot of discussion of consumers not wanting autonomous cars,” said Mark Wakefield, head of the automotive practice in the Americas at Alix. “That didn’t correlate with our experience.”
Still the timeline is long and drawn out: Alix predicts 2% market penetration of fully autonomous vehicles by 2025 and 7.5% by 2035.
Wants and Needs
In the UMTRI survey, a full 94% of consumers said that even in a fully autonomous vehicle, they wanted a steering wheel and gas and brakes. “We’ve explained, ‘You don’t need it,” said Schoettle. “And they say, ‘But I still want it.’”
Schoettle finds it worrisome that the public is more accepting of partial than full autonomy because throwing a driver into the mix creates additional safety concerns. Someone lulled into trusting the car to the point of texting or reading a book, then jolted into taking the wheel, can’t instantly size up the car’s predicament and make appropriate split-second decisions.
“You’re not starting from a dead stop,” says Schoettle of partially autonomous driving. “It’s not ‘Hey, it’s a beautiful day; I’m just going to take over and drive myself.’ You’re going 40 mph down the road and maybe the weather gets bad or something strange happens and the car says, ‘You need to take over.’”
Designing a fully autonomous car with the infrastructure for human backup, says Schoettle, is also not realistic from an engineering standpoint. The vehicle ends up with redundancies—like rearview mirrors and turn signals—when the computer has its own indicators and doesn’t need the human-centered ones.
“If you still have a steering wheel and brake, all of a sudden you need to have all of this stuff back in the car,” Schoettle said.
Cars vs. Horses
In a report last year on disruptive mobility, Barclays analyst Brian Johnson likened the timeline for autonomous vehicle acceptance to the time it took society to shift its dominant transportation mode from horses to automobiles. The peak year for the horse was 1920, a dozen years after the first Ford Model T rolled off the line.
“Being in the market by 2025 does not necessarily mean mass uptake,” says Shawn DuBravac, chief economist for Consumer Technology Association. “There’s a lot of moving parts. It could easily be another 20 to 25 years after the point that they first become available that we see a majority of vehicles sold to be autonomous.”
DuBravac predicts that around 2% of vehicles sold in 2025 will be fully autonomous, and they will be rolled out in the luxury market first, with an average price of around $80,000.
“You will see more acceptance as the cost comes down,” he says, noting that the LIDAR technology that used to cost $80,000 per vehicle now costs around $8,000.
But if the technology isn’t where it needs to be, all bets are off. Brian Daughtery, CTO of the Motor Equipment Manufacturers Association, struck a note of caution when he told an audience of automotive suppliers in Michigan in August that fully autonomous vehicles “are probably quite a bit further off than some of the articles that we see in the popular press would lead you to believe, expect in controlled environments.
“Humans have proven to be very adept at handling complex situations on the road, even if it’s something they’ve never seen before.”
Fully autonomous cars need more redundancy than traditional cars in case a system fails, he said. “Because of the driver, we essentially have a very low-cost method if an electric steering or braking system fails. You have a driver who’s able to perform those tasks manually and get the vehicle safely over to the side of the road and complete the journey as needed." Building a backup system into the car itself "gets expensive very quickly.”
A Marketing Challenge
Schoettle says that people need more of a chance to experience self-driving vehicles, or see others riding them around town in them, to wrap their brains around the concept. “Google did a great job in some of the places they are—Mountainview (Calif.) and Arizona—but you can’t do that in every single city.”
Word of mouth goes a long way, too. “Where enough people get to ride in it so when someone says, “I’m afraid it’s going to do this or that,” someone else can say, “Oh no, I used these and they’re great. They don’t do that.”
Schoettle thinks a good baby step for introducing fully autonomous technology is with vehicles for public use—the driverless Ubers planned for Pittsburgh, or the autonomous podcars used as public transit in Europe.
In those cases, “you’re not really being asked to give up control,” he said. “You’re just using it at the airport, or in the central business district. But asking people to do this with their personal vehicles is a big leap.”
He predicts that in 10 or 15 years, most people will know someone who drives a self-driving car—the guy down the street, for instance. “When people says this is 30, 50 years off, that strikes me as too long.”
“I think that experience—even if it’s just your cousin saying, 'I got one, you should check it out'—goes a long way.”