Safety concerns
On February 14, Google reported a minor traffic accident in which one of its self-driving cars struck a city bus in February in Silicon Valley.
"We clearly bear some responsibility, because if our car hadn't moved there wouldn't have been a collision," Google said in its monthly report in February about its self-driving car project.
Generally, AI makes for a much better driver than a human being, said Li Jun from PwC, noting that although more measures need to be adopted to prevent accidents, AI technology will mature as time goes on.
"For instance, automatic driving technology can kick in when a car gets into heavy traffic, which can reduce the frequency of traffic accidents," Li Xiang said, also noting that autonomous driving technology costs automakers less than $1,000, which is relatively low.
"Compared with full self-driving technology, I see a good future for autonomous driving technology over the next few years," Li Xiang told the Global Times on March 31.
Obstacles on the road
Self-driving technology developers still face many problems such as high costs and how to upgrade software systems, though they are all solvable, Li Xiang said.
Self-driving cars have to be able to sense their surroundings so they can decide how to safely maneuver on the road. That makes the type of sensors that developers use particularly important.
For instance, LIDAR (light-sensing radar) sensors, which Google uses in its self-driving car prototype, cost as much as $80,000, according to a report by U.S.-based tech information provider Tech Insider in December 2015.
Besides the high costs, there are many other factors that will decide whether the technology will be widely used in the future.
"The related transportation laws and regulations still need to be completed in China," Li Jun said.
The National Highway Traffic Safety Administration (NHTSA), the U.S.'s auto safety regulator, said the artificial intelligence system piloting Google's self-driving car could be considered a driver under federal law, Reuters reported on February.
The regulator's statement was a step toward winning approval to get autonomous vehicles on the roads legally.
In the e-mail, Baidu's Kuo pointed out that China's road infrastructure can't currently accommodate autonomous vehicles.
"They would have trouble, for instance, understanding the hand gestures of traffic cops or construction workers, would have difficulty understanding instructions shouted through a bullhorn or megaphone, would have difficulty distinguishing quickly between, say, a soft object like a mattress falling off the back of a truck or a hard object," he said.
AI has become a trend that will change people's driving habits, Li Jun said, noting that it's very hard for an individual company to dominate the competitive market.
Self-driving technology is expected to be tested in some areas during the 2022 Winter Olympic Games in Beijing, according to a report by domestic news website gd.qq.com on Friday.
About a decade from now, the total number of private vehicles in Beijing will fall to 1 million, said Wu Gansha, former director of the Intel China Research Center, according to a report on the news website huanqiu.com on March 31. By that time, however, Beijing will have about 2 million self-driving taxies on the road.
That prediction could be optimistic. It will still be several decades before self-driving cars are widely adopted, according to the Moody's report.
"In our projection, a majority of cars could be self-driving around 2045, and are likely to become nearly universal about 2055," the report said.
