How humanlike should self-driving cars be?
It’s a question that nuTonomy, a company that’s launched a self-driving taxi service in Singapore, is trying to answer. It appears that some version of the “uncanny valley”—a visceral negative response people feel to robots that seem almost human, but not human enough—also applies to automated vehicles.
“For better or worse, we have to bridge this divide between developing cars that drive by the book and cars that drive how you and I drive,” Karl Iagnemma, CEO of nuTonomy, said today at EmTech MIT 2016, in Cambridge, Massachusetts. “It’s an open question where on the continuum you want to drive—and it’s something we’re researching.”
Iagnemma says his company began offering rides in order to gather data about the public’s reaction to the technology. This effort, and others like it, will be vitally important, because public perception will play a big role in how rapidly automated driving is commercialized. “You can think of it as the world’s largest, most expensive focus group,” Iagnemma said. “It’s extremely valuable stuff, and there’s just no other way to get the information.”
NuTonomy is far from the only company working on self-driving cars, but it was the first to offer rides to members of the public, several weeks ago. Uber has since followed with a service available in Pittsburgh (see “Your Driverless Ride Is Arriving Now”).
Iagnemma says riders seem to experience a steep acceptance curve: a few minutes of nervousness quickly turns into complacency and eventually boredom. Still, they seem unnerved by driving behavior that strikes them as particularly unhuman. “Our first iteration of driverless cars kind of drove like trolleys on a track,” Iagnemma said. “This uncanny notion threw people off. We now appreciate that it’s vitally important.”
Perhaps the most high-profile trial of self-driving cars, being run by Google in several cities around the U.S., has also highlighted the importance of driving behavior. The unnatural way that the cars pull away from traffic lights has apparently resulted in some accidents (see “Google’s Self-Driving Car Probably Caused Its First Accident”).
Self-driving cars will also most likely need to adapt to driving in different cultures. A car trained on the roads of California, for instance, might prove too laid-back for the notoriously pushy streets of Boston or New York.
NuTonomy’s Singapore trial has highlighted additional issues that Uber and other automated taxis will face. The human side of taxi service—helping someone with luggage or returning a lost phone—will have to be worked around with self-driving cars, Iagnemma says.
The 50-year-old problem that eludes theoretical computer science
A solution to P vs NP could unlock countless computational problems—or keep them forever out of reach.
The moon didn’t die as early as we thought
Samples from China’s lunar lander could change everything we know about the moon’s volcanic record.
Forget dating apps: Here’s how the net’s newest matchmakers help you find love
Fed up with apps, people looking for romance are finding inspiration on Twitter, TikTok—and even email newsletters.
Inside the machine that saved Moore’s Law
The Dutch firm ASML spent $9 billion and 17 years developing a way to keep making denser computer chips.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.