China, second only to the United States, stands as America’s largest competitor, yet only 12% of the most elite AI talent prefer employment in China, highlighting a significant gap. Chinese AI talent remains highly active in the United States. For instance, approximately 20% of the core contributors to the GPT-4 team are from China.
Currently, 28% of top AI talent chooses to work in China, while 42% opt for the United States. It’s evident that in recent years, the gap between China and the United States has significantly narrowed, shifting from American dominance in the choice of AI talent workplace to a competitive dynamic between China and the United States.
Talent is the Key to Success
In the field of artificial intelligence, which profoundly influences humanity, talent, especially top-tier AI talent, is one of the crucial factors in winning this competition.
Recently, the Paulson Institute’s think tank MacroPolo released a survey called “Global AI Talent Tracking,” which outlines the movement trajectories of top AI talent, answering where top AI talent comes from and where they are headed.
MacroPolo utilized author data from the NeurIPS conference, recognized as one of the top conferences in the AI field. The research presented at NeurIPS focuses particularly on theoretical advancements in neural networks and deep learning, driving many of the latest advancements in artificial intelligence.
At the 2022 NeurIPS conference, 2,671 papers were accepted, with an acceptance rate of 25.6%. The opportunity to receive oral presentations at NeurIPS is even more prestigious. In 2022, only 190 papers received this honor, with an acceptance rate of just 1.8%. These authors, totaling 955 individuals, are identified by MacroPolo as the most elite members of the AI talent pool (approximately the top 2%).
Due to the large number of accepted papers, MacroPolo randomly selected 186 papers for analysis, with a total of 867 authors. Based on the authors’ educational backgrounds and institutional affiliations, MacroPolo summarized the characteristics and trends of the flow of top AI talent.
MacroPolo previously published a “Global AI Talent Tracking” survey in 2020, and this survey serves as an update to the database. Comparing the data from the two surveys reveals changes in talent flow trends.
Elite AI Talent Prefers Employment in the US, Followed by China
The data shows that China and the United States are the main sources and destination workplaces for top AI talent. 70% of top AI talent works in institutions in China or the United States, and 65% of top AI talent comes from China and the United States. Although other countries also have their characteristics, China and the United States still dominate the AI talent landscape.
In the competition between China and the United States, the United States maintains a significant advantage in top AI talent. The United States is home to 60% of the world’s top AI research institutions and is the preferred employment destination for the most elite (top 2%) AI talent globally, with 57% of the most elite AI talent preferring employment in the United States. China ranks second to the United States as its largest competitor, yet only 12% of the most elite AI talent prefer employment in China, highlighting a significant gap.
This advantage of the United States is largely due to its leading universities, particularly in graduate education, and the strong talent magnetism of Silicon Valley. With world-class institutions such as Stanford, MIT, and UC Berkeley, coupled with tech giants like Google and OpenAI and an active entrepreneurial ecosystem, the United States is in a favorable position to attract and nurture top AI talent.
In fact, the United States heavily relies on STEM fields and effectively utilizes foreign talent, built upon the attractiveness of its graduate education system to international students. Of the 33,759 STEM PhDs awarded in the United States in 2019, 39% were born outside the country. The number of STEM PhDs graduating annually in the United States from foreign students is nearly as high as the total number of STEM PhDs graduating in Germany each year.
China is the World’s Largest Exporter of Top-Tier AI Talent
The advantage of graduate education in the United States is fully reflected in the distribution of talent in the field of AI. China is the world’s largest exporter of top-tier AI talent and also the biggest competitor to the United States in terms of both quantity and quality of talent. Among the top (top 20%) AI talents who received undergraduate education in China, they account for 47% globally, while those who received undergraduate education in the United States only account for 18%.
However, in the graduate stage, a large number of Chinese AI talents flow to the United States, with nearly four out of ten Chinese AI talents choosing to pursue further studies in the United States, reversing the proportion of AI talents between China and the United States. After obtaining a doctoral degree in the United States, 77% of non-American students choose to stay and work in the United States.
Currently, among the top AI research institutions in the United States, the number of AI talents from China even exceeds that from the United States. Three-quarters of the top AI talents in these institutions come from the United States and China, with the talents from the United States and China accounting for 37% and 38%, respectively, with a slightly higher proportion of Chinese talents. In contrast, the source of top AI talents in China is much more singular, with 75% of the talents absorbed receiving undergraduate and graduate education domestically. Not only are foreign talents scarce, but also only about 10% of talents at the graduate level have experience studying in the United States.
From left to right, is undergraduate institution, graduate school, and post-graduate work location.
Chinese AI talents are very active in the United States. For example, approximately 20% of the researchers in the core contributors list of the GPT-4 team are from China. In terms of professional fields, the highest proportion of Chinese talents on the core contributors list of GPT-4 is in the field of computer vision, with 30% of the members coming from China. In the complete list of contributors to GPT4, there are 32 talents from China. The Chinese AI talents counted here do not include Chinese-American scientists born abroad.
Among these 32 Chinese AI talents who contributed to GPT-4, 11 completed their undergraduate studies in China, and 21 completed their undergraduate studies in the United States. At the graduate stage, close to eighty percent of these talents studied in the United States and stayed there. It can be said that receiving undergraduate education in China, studying for master’s and doctoral degrees in the United States, and working on top-level artificial intelligence research, this is the portrait of a generation of Chinese AI talents.
The GPT-4 team is not an exception in this regard. Specifically in the field of computer science, Chinese graduate students account for 14% of the total number of computer science graduate students in the United States, and more than a quarter of the total number of international students studying in the United States. Among them, 90% of Chinese graduate students choose to stay in the United States to develop their careers.
For the United States, the influx of Chinese AI talents has contributed to the rapid development and innovation of the United States in the field of AI. For China, these talents with backgrounds in graduate education from top universities in the United States and work experience in world-leading technology companies are the targets that governments and top technology companies hope to strive for. Their return will be of great help to the education and industry of AI in China.
The Gap between China and the United States is Narrowing
In fact, in recent years, China has made rapid progress in cultivating and attracting AI talents. According to statistics from the MacroPolo Institute, the proportion of top AI researchers cultivated by China in the world rose to 47% in 2022, a significant increase from 29% in 2019. The quality of AI talents from China has also improved. The most elite (top 2%) AI talents now account for 26% from China, while the United States is 28%, which is very close. In the previous survey, only 10% of these most elite AI talents came from China.
In terms of job choices, in the previous survey, only 11% of top AI talents (top 20%) chose to work in China, while the United States employed nearly 60% of the world’s top AI talents. However, this survey shows that now 28% of top AI talents choose to work in China, while 42% work in the United States. It can be seen that in recent years, the gap between China and the United States has narrowed significantly, and the choice of work location for AI talents has shifted from the dominance of the United States to competition between China and the United States.
This is largely due to the growing demand for talents brought about by the rapid development of China’s AI industry, as well as the increasing emphasis on the cultivation and introduction of AI talents by domestic universities and technology companies.
For example, in terms of research institutions, in the previous list of the top 25 AI research institutions compiled by the MacroPolo Institute (based on the affiliation of paper authors), Tsinghua University and Peking University were the only two Chinese research institutions. But in this year’s statistics, Chinese research institutions occupy six seats, and Shanghai Jiao Tong University, Zhejiang University, the Chinese Academy of Sciences, and Huawei have also entered the top 25. Tsinghua University and Peking University are ranked third and sixth, respectively.
More AI talents choose to receive graduate education in China rather than the United States. According to the MacroPolo Institute’s report in 2020, about 56% of AI researchers who received undergraduate education in China went to the United States for further studies. In this year’s report, this number is only about 36%. Since the country where graduate education is received is closely linked to the future work location, this also means that more talents will work in China in the future.
This may also be related to the inconvenience of studying abroad during the pandemic and the restrictions on studying in STEM fields in the United States for Chinese students. From 2019 to 2022, the total number of Chinese students studying computer science and mathematics in the United States decreased by more than 10,000. Chen Xi, co-founder and CEO of AI robotics company Covariant, said in an interview that further restrictions by the US government could harm the entire field, which would definitely affect the company’s ability to recruit talents. If this continues, his company may face difficulties in recruiting talents.
However, the problem cannot be solely attributed to the obstacles of the pandemic and the restrictions on Chinese students studying STEM fields in the United States, as India has also seen a similar trend of talent returning. Although India remains an important exporter of top AI talents, its ability to retain talents is increasing. In 2019, almost all Indian AI researchers (based on undergraduate degrees) chose to develop overseas, but by 2022, one-fifth of them had stayed in India for work.
The development trends of China and India seem to reflect a decline in the overall mobility of top AI researchers in recent years. In 2022, only 42% of top AI talents were foreign nationals working in other countries, a decrease of 13 percentage points from 2019, indicating that more talents are choosing to stay in their home countries. While the United States still maintains a clear advantage in top AI talents, other countries are also accumulating their own talent pools, challenging the dominant position of the United States.
