Robin Li, member of the national committee of the 13th Chinese People’s Political Consultative Conference (CPPCC) and Baidu
On the evening of March 2, Li conducted an interview regarding his proposals. This year he prepared a total of four proposals: national policy to encourage open AI platforms, government support for laws and infrastructures to develop unmanned driving, policy facilitation in Xiongan New District to attract talents, and wider tobacco control ordinance.
According to Li, in the AI era, an open platform will be highly influential. No single company can accumulate enough data, therefore, sharing data will make a larger impact. Baidu
Li also noted that at least 20 states in the U.S. have introduced laws and regulations on autonomous driving. In China, only Beijing and Shanghai have carried out relevant tests. The automotive industry is huge and involves many levels. Li believes that at all levels, the government can play a role in promoting autonomous driving. “The government does not need to build cars or invent autonomous driving technology. They may simply improve the infrastructure or facilitate something else to reduce difficulty of implementing autonomous driving technology,” said Li.
Robin Li was also asked more questions about the development of artificial intelligence. He answered questions regarding the advent of artificial intelligence, the comparison of AI development between China and the U.S., and the commercialization of AI.
AI has become a practical technique.
Li believes that technology has come to a tipping point. First, new AI technology is constantly developing; second, new data is constantly gathered; and third, computational power is constantly increasing. When these three aspects reaches a certain level, AI will indeed become a practical technology.
AI has been developing fast in recent years. The last few years have seen breakthroughs in speech recognition technology, image recognition technology, user persona, and personalized recommendations.
In Li’s view, AI will simplify complexities in daily life. “Today, we face traffic congestion, air pollution, lining up to buy tickets, go through security checks and authenticate identities. In my opinion, all of these all unnecessary.” He revealed that he was in talks with the Palace Museum for using AI to solve security problems.
Advantages of developing AI in China
Li said that the U.S. remains ahead of China in artificial intelligence, but China also shows three advantages.
First, data is a unique advantage for China. More than 700 million of China’s netizens speak Chinese and share Chinese culture and laws. Data produced by Chinese netizens is harmonious and unified. Valuable results can be achieved when using these data to train machine learning and build AI models.
Secondly, the Chinese government is relatively open and forward looking in terms of policies. The Chinese government is willing to allow trials and testing, even if the laws are not sound enough. The overall attitude is to encourage innovation. “For example, last year, I put unmanned vehicles on the Fifth Ring Road. The relevant departments did punish me for it, but overall speaking, the government tolerates innovation to certain degree,” Li said.
Third, China is relatively backwards in traditional industries, and thus have better acceptance of new technology than U.S. Li said, “Machine can make good predictions. The traditional industries, whether agriculture, industry, or services, are generally more open to AI. China has its own advantages in this aspect.”
AI is almost impossible to implement in the short term. AI requires large amounts of investment, and is still difficult to generate revenue from . Li agrees that the profitability of AI is indeed a problem. From larger enterprises like Baidu
Never felt that Baidu
Li said, “I never felt that I was a copycat, ever since the first day I established Baidu
“In fact, I went into the search engine area earlier than Google. I invented the hyperlink analysis, one of the basic inventions that decided development trends and direction of modern searching engines, earlier than Google. Technically, I do not feel we need to imitate others. We just study and analyze users to find out what the user really needs and how we can better meet those needs.”
“In fact, I think that in artificial intelligence, Google kind of mimics Apple in terms of having an end-to-end solution where everything from operating system to service is proprietary. The Baidu