According to the website of the World Artificial Intelligence Conference, at the 2023 World Artificial Intelligence Conference Industry Development Plenary Meeting held on July 7, the "2022 Global Artificial Intelligence Innovation Index Report" was officially released. The AI Innovation Index is an important indicator reflecting the national AI innovation level and was developed by the China Institute of Science and Technology Information. The report shows that the overall pattern of global artificial intelligence development led by China and the United States and distributed in a cascade remains unchanged. The artificial intelligence innovation index has maintained the second level in the world in the past three years, and has made progress in talents, education, patent output, etc., but the level of basic resource construction still needs to be improved.Chinese
First, the overall pattern of global artificial intelligence development led by China and the United States and distributed in a ladder will remain unchanged
According to the 2022 AI Innovation Index score ranking, the 46 participating countries can be divided into four major echelons. The first tier countries scored more than 50 points, with only the United States and China entering. The second-tier countries scored 35~50 points, including 11 countries such as the United Kingdom, Germany, and Singapore. The third-tier countries scored 20~35 points, including 12 countries such as Denmark and Finland. The fourth tier of countries scored less than 20 points, including 21 countries such as the Czech Republic and Brazil.
The United States' AI innovation index has ranked first in the world for four consecutive years, and China has maintained the second level in the world for three consecutive years. However, it is worth noting that in 2022, due to the addition of internationalization indicators, the gap between China and the United States has widened. Some countries in the other echelons have made significant progress, such as Sweden and the Netherlands entering the second echelon for the first time, and India, Poland, etc. entering the third echelon for the first time. Judging from the changes in the overall ranking in the past three years, among the top ranked countries, except for China and the United States, there is a trend of catching up between other countries.
Figure 1 Scores and rankings of AI innovation indices by country (2022)
Figure 2 Top 15 countries in the AI Innovation Index in the past three years
Second, the development of Chinese intelligence has achieved remarkable results, and the level of basic resource construction needs to be improved
Compared with 2021, China has 10 third-level indicators that have risen in 2022, mainly focusing on talent, education, patent output, innovation system and other aspects. The number of dominance indicators is also growing, from 15 in 2021 to 18 in 2022. However, there is still a certain gap between the overall level of development and the United States. Among the five first-level indicators, the United States ranked first, but China did not rank first; Among the 11 secondary indicators, China lags behind the United States in 9 indicators; Of the 33 third-tier indicators, China has 24 indicators that perform worse than the United States. In addition, compared with its own rapid growth in innovation output, the scale and quality of its innovation investment still have a lot of room for improvement. For example, the quality and openness of public data are not high, the relevant indicators rank low, and the information foundation still has a lot of room for improvement, ranking in the middle position of participating countries in indicators such as mobile cellular subscription rate, Internet usage, and fixed broadband subscription rate.
Figure 3 Comparison of the scores of secondary indicators in China and the United States
Third, global artificial intelligence is in a period of accelerated development
The index evaluation results in recent years show that the current global artificial intelligence is accelerating, showing several obvious trends.
First, strategic deployment and AI governance have received increasing attention. In terms of AI innovation system, the report focuses on two aspects: the completeness of AI policy planning in various countries and the importance of AI governance. Compared with 2021, almost all participating countries have increased their scores in 2022, whether it is the first-level indicator of AI innovation system or the second-level indicator of policy planning and governance. Among them, five countries, the United Kingdom, Singapore, the United States, China and Australia, improved their policy planning scores the most, and the United Kingdom, Singapore, France, Australia and Japan improved their governance scores the most.
Second, the construction of data and computing power infrastructure continues to advance. In terms of the number of data centers, the total number of colocation data centers in the participating countries has continued to grow in the past three years, and one-third of them have increased by more than 10% compared with 2020. From the perspective of the number of supercomputing, as more and more countries increase the construction of computing power infrastructure, the global supercomputing top 500 ranking is fiercely contested, and China's advantage in supercomputing has weakened. From 2020 to 2022, the number of supercomputing companies in China entering the Global 500 decreased year by year, and its proportion in the world decreased from 45% in 2020 to 35% in 2022. The number of supercomputing companies in the United States, the United Kingdom, Germany and other countries entering the global top 500 continues to increase.
Third, the industrialization of artificial intelligence has accelerated. In 2022, the total number of AI companies and the total number of AI employees in the participating countries will continue to grow, and the growth rate will expand. The total number of AI companies increased by 25% year-on-year, up from 18% in 2021; The total number of AI employees increased by 53% year-on-year, significantly higher than the 10% in 2021.
Fourth, artificial intelligence accelerates scientific research. In addition to information technology topics such as electronics, communication, and computer science, artificial intelligence papers also involve basic disciplines such as environmental science, geographical science, and material science. From 2020 to 2022, the number of AI papers involving basic disciplines such as environmental science, geographic science, and materials science has continued to grow, accounting for 5% of the total number of AI papers in 2020 to 10% in 2022.