[Text/Observer Network columnist Liu Dian]
The explosion of ChatGPT has focused the world's attention on generative artificial intelligence and its future applications. In the field of Chinese artificial intelligence, artificial intelligence companies such as Baidu, 360, and iFLYTEK have made efforts, and Huawei has also released words that "there has been a layout in this regard." There is a lot of heated discussion in the industry about who will eventually spend the "Chinese version of ChatGPT", but when we look far, we will find that the question is not just about who will become the industry hegemon.
Why is the AI field in the United States first?
In October 2021, Nicholas Sharon, the first chief software officer of the US Air Force, said in an interview with the Financial Times that China has taken the lead in artificial intelligence technology for the United States and is moving towards global dominance.
This remark caused an uproar in China, when everyone thought that China had achieved industry leadership in the field of artificial intelligence. It was not until last year's overnight explosion of OpenAI company ChatGPT that we understood that this remark was just a "war in the United States", and the artificial intelligence competition between China and the United States was far from victory.
GPT (Generative Pre-trained Transformer) in the name ChatGPT is a deep learning model launched by OpenAI. The Transformer included in its name is a model first proposed by the Google Brain team in the 2017 paper "Attention is all you need", which is a milestone in the field of artificial intelligence. So far, many landmark basic theories and algorithm research in the field of artificial intelligence have been first proposed by the United States. ChatGPT is the result of building on this series.
The United States can take the lead in artificial intelligence technology such as ChatGPT, first of all, because of its technical reserves and layout advantages. The United States is the birthplace of artificial intelligence, and many universities and enterprises have contributed a deep theoretical and algorithmic foundation to the development of artificial intelligence. Its artificial intelligence technology reserves and layout foundation are deep, and it has always been in a global leading position.
At the basic layer of the industry, the strength of chips and sensors in the United States is strong, and the representative manufacturers include NVIDIA, Qualcomm and other leading manufacturers in the industry with strong technical strength. Algorithm research in the fields of computer vision and speech recognition in the United States has also begun as early as the sixties and seventies of the last century, and startups in these fields have emerged in recent years.
In contrast, China has accumulated a lot of strength in the artificial intelligence technology layer and application layer, but its overall strength in the basic layer is still weak. Especially in the field of chips, China is often "stuck in the neck" in key semiconductor production equipment and electronic design automation.
The gap between the two countries in the basic, technical and application layers of artificial intelligence is also reflected in the technology giants of the two countries. American manufacturers such as IBM, Microsoft, Google, Meta, Amazon and other large manufacturers have a comprehensive layout in algorithms, computing power, data and other technologies. Baidu, Alibaba, Tencent and other domestic manufacturers have layouts in core technology fields such as computer vision and natural language processing. And due to the rich application scenarios in China, Chinese giants are ahead in terms of application and commercialization, but the basic research part is still lagging behind.
At present, the United States is increasing research investment in cutting-edge fields such as neuromorphology, optical computing, and quantum machine learning, and measures such as allocating more research funds by the government and enterprises and relaxing the immigration conditions for high-tech talents to accelerate breakthroughs in cutting-edge technologies. At the same time, it will increase the technological blockade of China. In October 2022, the U.S. government updated and expanded restrictions on the export of technology and chips to further halt China's progress in developing cutting-edge technologies.
For decades, the United States has attracted top global talent in the field of artificial intelligence, dictated by its national culture and immigrant traditions, and although China has also increased its attention in this area in the past two decades, there is still a clear gap with the United States. According to an article by Stanford HAI, the hiring rate of AI workers in the United States in 2020 was about twice that of 2016. By comparison, China's growth rate over the same period was around 30%. According to think tank MacroPolo, while China has produced a large number of top AI researchers, most of them continue to live, work and study in the United States.
Bringing together more talent has two important aspects. One is to cultivate more AI talents within the country's education and social systems. The other is to create more opportunities for teaching and markets in the country and attract foreign talents. In terms of training, the United States is represented by Carnegie Mellon University, Stanford University, Massachusetts Institute of Technology, and Google to cultivate the world's top artificial intelligence talents, while China is represented by Tsinghua University, Peking University, Microsoft Research, and the Chinese Academy of Sciences.
In the United States, high-quality AI research comes not only from academia, but also from companies. Much of China's research comes exclusively from academia. In terms of attraction, due to the advantages of employment and salary in the United States, it can attract top researchers from all over the world to engage in research and work in the United States, and in recent years, under many factors such as China's salary increase and the incentive policy for international students to return to China, talents have also shown an upward trend.
In addition to academic research and development, the development of artificial intelligence also depends on the development of industry. A strong industry can support the research of cutting-edge technologies, and breakthroughs in new technologies can in turn give new vitality to the industry. Therefore, both China and the United States have invested heavily in the artificial intelligence industry and carried out forward-looking layout early.
In contrast, China's early start in industrial policy and support is strong, which also gives Americans the illusion of "rising strongly". China's State Council launched the "New Generation of Artificial Intelligence Development Plan" in 2017, which for the first time explicitly elevated AI to a national-level strategy, emphasizing the major impact of AI on international competition, economic development and social governance, and then various departments successively issued specific support and funding projects. Although the United States launched "Preparing for the Future of Artificial Intelligence" during the Obama era, emphasizing its importance in helping the United States maintain global technological leadership and ensure national security, the real policy was introduced much later.
In 2019, US President Trump signed an executive order to launch the American Artificial Intelligence Initiative, with three keywords: "top-level promotion", "American leadership" and "focus on the foundation". Prior to this, the White House established a "special committee on artificial intelligence", and the Defense Advanced Research Projects Agency, a US military research organization, announced at the same time that it plans to invest $2 billion in the next five years to develop the next generation of artificial intelligence technology to promote the "third wave" of machine learning.
In 2022, the Biden administration signed the Chip and Science Act of 2022 and the Inflation Reduction Act of 2022, both of which mentioned the deployment of the US industrial chain and supply chain security. It is hoped that through government intervention and stimulating industrial policies, more high-tech manufacturing industries originally deployed overseas will be relocated to China, including chip manufacturing closely related to artificial intelligence technology. The United States' industrial policy on cutting-edge technologies such as artificial intelligence has gradually strengthened, partly out of the need to compete with China, and under the stimulation of relevant industrial policies, the development of the artificial intelligence industry in the United States has also accelerated significantly.
The big country technology power game behind AIGC
Although the competition in the field of artificial intelligence has just begun, the future is still unknown, but artificial intelligence technology is related to national security and great power competition, and it is a technology train that we cannot miss. The generative artificial intelligence where ChatGPT is located is directly linked to important applications such as search engines on a global scale. If any Chinese or U.S. company becomes a giant of global search engines because of the technological superiority of generative AI, the geopolitical implications will be far-reaching. Baidu announced that it will launch its AI chatbot in March 2023, which functions similarly to ChatGPT. In this field of generative artificial intelligence alone, language models will become one of the sub-battlegrounds of Sino-US technological competition.
Over the past hundred years, the United States has established global dominance in military, commerce, finance, and technology with unparalleled technological superiority. Therefore, the United States has made every effort to prevent the flow of top technology from countries around the world to China, and to prevent China's technology companies and products and services from going global. The core goal of the geopolitical game is to look at the game of power between countries from a geographical perspective. The reason why the field of artificial intelligence can become the target of power games lies in the power meaning behind artificial intelligence technology.
At the economic level, artificial intelligence technology will greatly improve labor productivity, promote economic development and productivity improvement. From an economic perspective, AI may be the next "communications revolution", promoting a huge increase in productivity and leading the next economic cycle.
Based on a study of 12 developed countries, Accenture Research and Frontier Economics predict that AI could "double the annual economic growth rate" by 2035 while also increasing labor productivity to 40 percent. The McKinsey Global Institute predicts that AI could generate $13 trillion in revenue from global economic activity by 2030. PricewaterhouseCoopers puts forward a higher figure, with global GDP growth reaching $15.7 trillion by 2030, much of which will be attributed to productivity gains. The predictions of these world-class institutions all point to the conclusion that artificial intelligence will achieve huge economic growth on a global scale.
At the same time, countries with leading AI technologies will once again be able to reap uneven economic gains from technological dominance. The McKinsey Global Institute study notes that "AI leaders, primarily in developed countries, can increase their lead over developing countries" and that "leading AI countries could reap an additional 20 to 25 percent net economic benefit compared to today, while developing countries may only account for about 5 to 15 percent." The logic goes back to the distribution of global value chains, where countries at the top of the value chain are more able to derive greater economic and monopoly benefits from high-value industries.
At the military level, artificial intelligence technology will directly act on the modernization and future of the military to improve military strength. The application fields of artificial intelligence include combat systems, strategic decision-making, data processing and research, combat simulation, target recognition, threat monitoring, drone swarms, network security, etc. Compared with traditional systems, artificial intelligence military systems can better manage massive amounts of data, enhancing the self-regulation, self-control and self-driving capabilities of combat systems. Generative AI for civilian functions such as ChatGPT can also be used directly for military purposes, such as generating military reports, translation, or communication roles. At present, countries including China, the United States, Russia, the United Kingdom and Israel have begun to apply AI technology in different aspects of their militaries to improve operational, logistical and training capabilities.
At the governance level, the formulation of global standards and technical specifications for artificial intelligence is still in a vacuum. International rule-making and governance may seem less practical than economic and military power, but these invisible norms often contain greater power. On the one hand, the dominance of international norms means that who can decide the "right way" and rules, then the party that violates international norms will be in a moral and public opinion inferior position, and will be more easily isolated by international partners. On the other hand, the dominance of international norms also means that the international community recognizes a certain party to a great extent, not only to recognize its past achievements and contributions, but also to recognize the culture, ideology and values behind the technical norms, and to give the dominant party great convenience in developing technology.
Global AI rankings
The future of AI, what is China?
As mentioned earlier, ChatGPT and its generative artificial intelligence are only part of the current competition in the field of artificial intelligence. On the one hand, if artificial intelligence is indeed the core of the next technological revolution, who can stand at the forefront of technology and take the lead in building the core technology layer and industrial chain, who will win the technical, economic, military and governance advantages with certain barriers. On the other hand, if the role of artificial intelligence is far from that great, or the era of artificial intelligence is still early, then the technology of gradually improving artificial intelligence can also transform various parts of economic and social life and the military.
At present, the generative artificial intelligence triggered by ChatGPT is pushing the artificial intelligence competition between China and the United States to a small climax. How is China responding to the AI technology revolution?
First of all, we must not be fooled by the "North American war", and we must be aware of our shortcomings.
On the one hand, China's weakness in basic research and algorithms of artificial intelligence has no shortcuts and detours. At present, the top papers and major theoretical innovations in the world are still dominated by the United States, the United Kingdom, Canada and other countries. Therefore, not only the government and universities need to pay attention to the importance of basic research, but enterprises and the market must pay more attention and patience in this regard.
On the other hand, key hardware aspects such as chips often encounter the danger of "stuck neck". At present, the upstream of international high-end chip manufacturing and supply chain is monopolized by international giants such as NVIDIA, Qualcomm, and Intel, and the United States is stepping up the management of international giants and affiliated enterprises, and is bound to cut off any way to flow to China's high-end chips. Therefore, technological autonomy and chip autonomy are the primary issues we must face at present.
Since the release of ChatGPT, Chinese technology giants have also revealed that they are engaged in this research. Baidu said it would complete internal testing of its large-scale artificial intelligence model, the Ernie Bot, in March. The Damo Academy, an Alibaba research institute, is also testing a ChatGPT-like tool. Overall, China is accelerating the development of the artificial intelligence industry.
Two cities at the forefront of Chinese intelligence, Beijing and Shanghai, have introduced support measures. Beijing is currently the city with the highest concentration of Chinese intelligent enterprises and talents, with 1,048 core artificial intelligence companies. Beijing announced that it will support leading enterprises to build large-scale artificial intelligence models, support key enterprises to invest in the construction of open source frameworks, and accelerate the supply of basic data. Shanghai has identified artificial intelligence, biomedicine and chip development as the three "pillar" industries for the city's long-term development.
At the same time, we should also recognize that China has two advantages in developing the artificial intelligence industry that cannot be ignored.
One is China's institutional advantage of concentrating its efforts to do big things. China's top-down decision-making and centralization can help us develop the AI industry faster. In recent years, China has successively issued a number of policies to encourage the development and innovation of the artificial intelligence industry, such as the Guiding Opinions on Accelerating Scenario Innovation and Promoting High-quality Economic Development with High-level Application of Artificial Intelligence and the Three-Year Action Plan for the Development of New Data Centers (2021-2023). In addition to providing support funds for enterprises, the government is also actively seeking markets for enterprises to promote the sustainable development of the artificial intelligence industry. On the other hand, China has a broad market advantage. Many real industries in mainland China have strong demand for artificial intelligence, and at the same time, the broad market also provides the development of artificial intelligence and its rich application scenarios.
Taking the generative artificial intelligence to which ChatGPT belongs as an example, China should rely on its own advantages to promote the blooming of upstream and downstream players in the industrial chain. Generative AI can be divided into application layer, model layer, cloud computing platform, and computing hardware layer. Therefore, the upstream of generative artificial intelligence includes data providers, algorithm institutions, etc.; Midstream includes text and image processing manufacturers; Downstream are content creation and service platforms and institutions. On the one hand, downstream institutions can quickly commercialize the industry with the advantages of broad market and rich application scenarios to meet social needs. On the other hand, the state and all kinds of enterprises should also polish their technical advantages at the model level, improve machine learning models, and enhance the industry competitiveness of Chinese manufacturers.
At the Beijing Artificial Intelligence Industry Innovation and Development Conference held on February 13, the Beijing Municipal Bureau of Economy and Information Technology officially released the "2022 Beijing Artificial Intelligence Industry Development White Paper". According to the white paper, in 2023, Beijing will continue to promote the construction of an AI innovation source with global influence, consolidate the foundation for the development of the AI industry, and build a high-level AI talent highland. Beijing is a microcosm of the current pace of Chinese development of industrial intelligence: China is accelerating its development in all aspects such as basic research and cutting-edge technology, talent training and attraction, industrial support and regulatory policies.
Although China failed to take the lead in launching ChatGPT, in the field of artificial intelligence big models, the massive data of Chinese content still has the potential to become high-quality and creative content output. For researchers, there must be a certain sense of urgency, but for investors, it is necessary to calm down and allow the slow commercialization process of basic research and technology.
This article is an exclusive manuscript of Observer.com, the content of the article is purely the author's personal opinion, does not represent the platform's views, unauthorized reproduction, otherwise legal responsibility will be pursued.