Source: Hong Kong Ming Pao

Author: Huang Weihao

China and the United States in the increasingly severe artificial intelligence (AI) competition, fighting is difficult to distinguish.The interesting part of this competition is that only a few years ago, China was regarded as a leading and better competitive advantage; however, in the era of generating artificial intelligence (Generative AI), the original advantages were weakened at any time to form a watershed.

Two different development directions of technology

Today, the development of AI has two development directions that can be completely different from the previous Internet -control (Control) or liberal.It can be considered that the name is mostly related to the country, and the government's governance and control are strengthened from the top to bottom. The latter is closely related to the market and the citizens.Civil and private enterprises have the performance and application of science and technology to promote development and popularization, and meet the needs and desires of all aspects of the entire society.

Of course, the two development directions may not be completely excluded from each other in theory; but in reality, it is not easy to achieve both fish and bear's paw.In the AI era, at the same time, the strong government emphasized the former, as well as the large market and active civil society advocated by the latter, is not a simple matter.

Having said that, in a time, China was able to take into account both functions in AI development at the same time and drive together. Therefore, it made the United States feel that its status of a large scientific and technological country is being threatened by unprecedented threats.At the government level, AI technology such as face recognition and other technologies, in conjunction with the "Social Credit System", has enabled the government to master the unprecedented information advantages.

In terms of market and society, through private enterprises such as Alibaba and Tencent, which make AI and big data such as technology, large -scale use in the category of production, shopping, entertainment, etc.China has quickly entered the "Digital Society".As early as many years ago, the author heard a lot of "complaints" in the Mainland students who went to study in Hong Kong. When the Mainland had generally used electronic payment, it was expensive to be Hong Kong in international cities and the world financial centers.Back behind technology.

"Age of Invention" and "Age of Execution"

Under this advantage of China, internationally renowned AI experts and entrepreneur Li Kaifu, he proposed an important prophecy in his influential book AI Superpowers, that is, China will defeat the United States in this AI competition.

Li Kaifu's conclusion is based on its reasonable analysis and obvious evidence.First of all, his statement is to build an important argument or assumption, that is, AI technology will not have significant breakthroughs in the future. The successful development of human distance can master and use knowledge.EssenceTherefore, he believes that AI has officially entered an "Age of Implementation" by an "Age of Discovery".

He believes that although the United States can lead in the "age of invention", the Chinese country has a better advantage in the "era of execution". This is because in the two levels of the government and the market mentioned earlierThe United States leads.At the government level, the source of the wisdom of AI is data, and because China has always been a large data country, naturally it has an advantage in "training" AI more than any country, making it smarter, prediction and more accurate.In the market and folks, a group of entrepreneurs who dare to use technology to become rich in technology and consumers who do not resist science and technology and trust the government have made China even more powerful in the AI competition.

Chatgpt Born to reverse the situation

But in just a few years, major changes have occurred throughout the situation, and even the possibility of 180 degrees reversal.Li Kaifu's AI Superpowers was published in 2018, only about 5 years before, but in the report of the Economist Magazine in May this year (Just how good can China get at Generation AI? "Essence

The biggest difference between the analysis of economists and Li Kaifu is that the former believes that the generation AI has risen rapidly, including the birth of ChatGPT, which has reversed the entire situation and subverted the development ecosystem of the entire AI.Due to the potential and value of the genetic AI, the investment and research of it has risen at the speed of explosive.In other words, the AI stage was returned to the original "age of invention."

When execution has always been a strong point in China, and invention and creativity are usually the traditional American advantages. Li Kaifu himself agreed in the book that this is a weaker link in China.China's powerful and powerful government is conducive to implementing decision -making, which is obviously China's institutional advantage; but in the AI "invention of invention", the chaotic situation of "no king's tube" often creates a situation of hundreds of flowers and hundreds of competitions.Assist endless creativity.

In fact, in terms of high -tech supervision, the United States has adopted a passive attitude since the emergence of the Internet.The US position is "Do NOT FIX What is Not Broken", which emphasizes the market's self -restraint and supervision. I am afraid that any supervision may hinder the market's creativity and reduce corporate business interests and return.Even more than a thousand people such as the science and technology giant Musk issued an open letter, asking for a half -year study of the suspension of generating AI, and the U.S. Congress also opened a hearing. Everything was only heard that the stairs were rang, and the thunder was small.

According to the analysis of economists, the Sino -US AI war has also been eliminated.When the situation turns sharply and is conducive to the strength of the United States itself, China's original advantages also showed signs of weakening.Unlike the United States, Chinese technology companies are facing more supervision of the country.In terms of data, although China has massive data, not all data is opened to the market and enterprises to do AI research at the same time.

Studying AI in China must consider the "emergence"

The consideration of national security has also become an additional consideration of generating AI research and its product development in China.Because of the characteristics of technology such as ChatGPT and other technologies, the computer absorbs data by itself to perform "Machine Learning" and gets the answer. ThereforeFunction.Technically, this is called "Emerget Properties".

Under the premise of Guoan, who should be responsible if AI "committing the law"?I believe that this is ChatGPT and many foreign and Chinese -generated AI companies. For the addition of how to launch its products in China, it is also the challenge that China needs to defeat the United States in the AI competition.

(The author is Academician of Study Science Higher Research Center of Stanford University)