By Yang Xite
Since the beginning of the Donald Trump era, the United States has gradually increased its restrictions and blockades on China in the field of high technology, seeking to stem the development pace of the latter’s advanced technology and maintain its advantage. This approach has continued unchanged into the current Joe Biden era. The same is true for the field of artificial intelligence (AI). However, within the field of AI, the U.S. seems to also have a considerable degree of willingness to cooperate with China.
From last year to now, at various levels of exchanges between the American and the Chinese governments, the U.S. has consistently expressed its willingness to cooperate with China on the global AI regulatory framework. The most obvious example is that at the inaugural global AI Safety Summit in November last year, the Chinese delegation was officially invited to attend and deliver speeches. Considering the close relationship between the UK and the US, it is hard to imagine the attendance of the Chinese delegation without US consent. Recently, high-level officials from China and the US have engaged in a dialogue on artificial intelligence in Switzerland. Suddenly, China seems to have returned to the global stage in the field of AI. However, based on comprehensive analysis, researchers at the Anbound Think Tank believe that the technological cold war between China and the US will only continue and intensify, with no possibility of any real turnaround, especially in the field of AI.
Firstly, the importance of AI technology dictates that the U.S. must exert full pressure on China. Since the birth of modern science and technology, obtaining comprehensive and efficient support from technology in production has always been a problem that scientific research has sought to solve. Whether it is the first three industrial revolutions or the fourth technological revolution, new technologies have not completely liberated people from tedious and repetitive monotonous labor and enabled them to engage fully in truly creative work. The rapid advancement of AI technology, especially the emergence of generative AI technology, first makes this prospect a realistic possibility. Existing research indicates that through automation and intelligence, AI can help us complete tedious and repetitive tasks, freeing up more time and energy to engage in more challenging and creative work. It can automate production processes, optimize supply chain management, improve product design, and promote the improvement of work efficiency and innovation capabilities.
In addition, the significance of AI lies in its application in scientific research and technological innovation by assisting scientists in large-scale data analysis and simulation experiments, thereby accelerating the research process. It can help discover new materials, develop new drugs, solve complex scientific problems, and promote the progress of science and technology. The prospects for AI to enhance military efficiency and military strike capabilities are also considerable. As it stands, the transformation of AI technology will be disruptive, and the country that leads in this technology will become a true hegemon in the global political and economic arena. For China, which has already been defined as its systemic competitor, the U.S. naturally cannot allow China’s AI technology to develop comprehensively and must deeply suppress it.
Secondly, the development trend of China’s AI industry and technology in recent years indicates that the U.S.’s suppression measures need to be further strengthened. Recently, a report released by the MacroPolo, the in-house think tank of the Paulson Institute, shows that in the global AI talent field, the U.S. remains the undisputed global hegemon, but China’s strength in talent cultivation and talent growth prospects is rather strong.
In 2022, the proportion of top AI researchers trained in China globally rose to 47%, a significant increase from 29% in 2019. The quality of AI talent from China has also improved, with the most elite (top 2%) AI talent now accounting for 26%, compared to 28% from the U.S., showing a very close margin between the two. In the previous survey in 2019, only 10% of these elite AI talents were from China.
Regarding job choices, in the previous survey, only 11% of top AI talents (top 20%) chose to work in China, while the U.S. hired nearly 60% of them. However, this survey shows that now 28% of top AI talents choose to work in China, while 42% work in the U.S. It can be seen that in recent years, the gap between the two countries has narrowed significantly, with the trend of AI talent job location shifting from U.S. dominance to competition between the U.S. and China.
In the previous list of the top 25 AI research institutions compiled by the MacroPolo (based on the affiliations of paper authors), Tsinghua University and Peking University were the only two Chinese institutions. But in this year’s statistics, Chinese research institutions occupy 6 seats, with Shanghai Jiao Tong University, Zhejiang University, Chinese Academy of Sciences, and Huawei also entering the top 25, with Tsinghua University and Peking University ranking third and sixth respectively. Previously, about 56% of AI researchers who received undergraduate education in China chose to further their studies in the U.S. However, in this year’s report, this number is only around 36%. Since the country where the graduate education is obtained is closely related to future job locations, this also implies that more talent may work in China in the future.
The report suggests that these trends are largely due to the increasing demand for talent brought about by the rapid development of China’s AI industry, as well as the rise of emphasis on talent cultivation and introduction by Chinese universities and technology companies in AI. All in all, the increasingly strong indigenous AI technology strength in China is the real reason for the change in the trend of top AI talent flow. Although the research strength and global attractiveness of the AI field in the U.S. still rank first globally, the impact from China implies that the U.S. needs to take further measures in regard to the development of China’s industry and research, while further strengthening the American research attractiveness.
When it comes to hardware, Huawei successfully released a 7-nanometer chip smartphone last year. Although in-depth analysis suggests that the main driver behind this breakthrough was the utilization of ASML lithography machines imported during the gap period of sanctions, the continued advancement of China’s research and development capabilities is also a reality even under multiple rounds of sanctions. Therefore, the U.S. will need to further strengthen its efforts to more effectively curb the progress of Chinese AI technology.
Furthermore, for a long time, the American business community has been an important opposing force in U.S. sanctions related to China. Given the substantial investment and profit demands of the business community in China, there has been consistent opposition to deep and comprehensive sanctions against China.
Taking NVIDIA, a heavyweight in the AI field, as an example, its CEO, Jensen Huang has repeatedly stated that the Chinese market is crucial, even pivotal, and NVIDIA will not leave the Chinese market. After the U.S. escalated sanctions, it still had to maintain exports to China even if it meant reducing the chip functionality. However, data disclosed in the fiscal years 2023-2024 show that although Mainland China contributed to 21% of NVIDIA’s total annual revenue, it had dropped to single digits in the fourth fiscal quarter, and its proportion may continue to decline. However, the decline in Mainland China revenue did not hinder NVIDIA’s overall revenue from continuing to strengthen throughout the fiscal year. Therefore, for leading American AI companies, the importance of the Chinese market is steadily weakening. Moreover, China’s recent call for increased utilization of domestically produced chips will undoubtedly further undermine the confidence of foreign AI companies in China.
For the entire U.S. business community, the importance of the Chinese market is also decreasing. Last year, American and Western companies withdrew hundreds of billions of dollars in profits from the Chinese market, and investment institutions continue to reduce investment scales. This also means that once the government is ready to launch corresponding deep sanctions against China, the pressure from the business community to oppose them will be even more limited.
According to previous reports from The New York Times (NYT), more than 100 tech leaders from Silicon Valley are lobbying Congress on the field of AI, urging Congress to impose more restrictions on AI from Mainland China. Palantir CEO Alex Karp and Roelof Botha, the managing partner of the venture capital firm Sequoia Capital, were also present. Palantir is a data analysis and AI technology company and a significant defense contractor for the U.S. government, rumored to have assisted the U.S. military in locating and targeting al-Qaeda leader bin Laden. It is reported that Mike Gallagher, who announced his resignation as a congressman earlier, will join Palantir. The NYT stated that this lobbying force from Silicon Valley not only aims to urge Congress to restrict China’s progress in the AI field but also hopes to improve related AI regulations, expand technology research and development spending, and relax immigration restrictions to attract more multinational AI talents to the U.S. These lobbying efforts are undoubtedly aiming to continue containing China’s progress in technology and increase the attractiveness of talent to China, thereby striking a blow to the country’s AI technology development.
ANBOUND’s founder Kung Chan has previously pointed out that in the field of AI systems, China’s restrictions are excessive and overly crowded, which is very unfavorable for the research and application of technology. Its overall level of openness and social freedom is also unfavorable for the further improvement of AI technology. At present, China’s AI technology development status mainly benefits from the accumulation of the past few years, mainly in a stock status. However, with the further deepening and strictness of U.S. sanctions measures, as well as the strengthening of targeted measures, the development of technology and hardware is restricted, coupled with the further highlighting of the trend of talent outflow, China’s AI industry development is likely to become even more challenging.
Final analysis conclusion:
As a global technological trend, while the U.S. and China may continue to cooperate in AI regulation, this will not alter the trajectory of the former tightening its control over AI against the latter. As it stands, geopolitical dynamics dictate this reality. During the Cold War, despite cooperation between the U.S. and the Soviet Union to prevent nuclear proliferation, the U.S. aggressively countered Soviet advances in nuclear technology. Similarly, in the ongoing technological rivalry between U.S. and China, some liken the American approach to “small yard, high fence”. Looking ahead, the yard may not get smaller, but the fence is about to get even higher.
- Yang Xite is a researcher at ANBOUND