Safety concerns, AI's limitations delay widespread adoption of self-driving vehicles
Will self-driving vehicles ever become commonplace? If so, when? What obstacles stand in the way?
To answer the first question: Yes, according to experts who took part in a panel discussion Friday at the University of Notre Dame’s IDEA Center, part of its Innovation Park complex.
Self-driving vehicles ultimately will be here. How quickly that happens remains a matter of debate.
There was wide agreement among the panelists that self-driving vehicles will first take hold among commercial fleet vehicles, such as ridesharing apps like Uber. And such uses will first roll out in the southwest, where weather conditions are more predictable than here in the Midwest.
Panelist Kevin Gay is head of safety for autonomous mobility and delivery at Uber.
“We expect, probably over the next three to five years, to see somewhere maybe around 10 to 20% of trips on the Uber platform in an autonomous vehicle,” Gay said.
The first self-driving Uber vehicles are planned for launch later this year in Phoenix, Gay said. He said simulation technology, training the vehicle’s systems to recognize potential crashes, still needs a lot of work.
Lawrence Burns, a former General Motors executive who now consults for the self-driving vehicle industry, said widespread adoption is taking longer than initially expected. He started working on the concept with Google in 2009.
Burns said the U.S. is leading the race to develop self-driving vehicles but China, which is ahead on electric vehicles, isn’t far behind.
China already leads the world in solar energy and electric vehicles.
“I think we always lead as a nation on the scientific and the early prototyping and the big idea,” Burns said. “But then when it comes to getting it to scale and getting it to learn… you see, what engineers do for a living is we make what’s possible real, and I think China just seems to have a faster system for doing that, once they believe something’s possible.”
Panelist Naseeb Souweidane is a Detroit-based researcher and advocate for U.S. automakers. He agreed China has an advantage in being more quickly able to build public support for self-driving vehicles. In China, the state controls news media, so there is much less transparency around injury accidents with self-driving vehicles that occur during research and testing.
For example, in 2018 when a self-driving Uber killed a pedestrian walking across the street in Tempe, Ariz., the accident drew wide national press coverage and an investigation by the National Transportation Safety Board. The federal agency has published its detailed findings online.
Souweidane said that kind of transparency is ultimately best for Americans.
“China is going ahead and pushing through these vehicles in ways that we don’t have that here, and it’s probably for the best that we have more confidence throughout our public in riding these vehicles later on,” Souweidane said.
In the U.S., Souweidane said how soon self-driving vehicles go mainstream will depend on buy-in from consumers. There’s wide agreement among experts that self-driving vehicles will be safer because they’ll eliminate human error.
“But if you just tell everybody they have to trust you that it’s going to reduce accidents by 20,000, I don’t believe that the public will just fully buy in, and pay more for these vehicles, ride them and go for it,” Souweidane said. “The perception of them has been going down each year. When these accidents happen it’s not really the technology’s fault but it’s misuse, so once we become more sophisticated with it, I think it will really take off and we’ll see more lives saved.”