It's not just the tech powers that are vying for the value of AI, countries around the world are vying for the leadership of AI, which is the emerging geography of AI. On December 12, 2023, Harvard Business Review (HBR) released the report "Charting the Emerging Geography of Al". The data is reported from more than 20 different institutions, including public databases such as the International Telecommunication Union (ITU) and the World BankProprietary databases such as the Center for Data Governance at George Washington University in the United States;and Harvard Business Review's own database and Digital Planet model, which assessed AI leadership in 25 countries and identified 4 key factors that make countries leaders in AI. Meta-strategy compiles important contents of the report and provides reference for how countries can maintain their leading position in the field of AI.
Determines AI leadership.
Key factors and evaluation indices.
The rapidly accumulating data pool in the global digital economy is clearly one of the key drivers of AI development. In 2019, HBR introduced the concept of a country's "gross data product," which is determined by the volume, complexity, and accessibility of data consumption, as well as the number of active internet users in the country. In this analysis, HBR recognizes that data production is an important asset for AI development, especially for generative AI that requires massive and diverse datasets. HBR has updated its 2019 analysis as a basis for its research, adding new drivers that are critical to the overall development of AI. To compare countries' AI leadership, the report identifies 4 key factors that determine a country's leadership in the AI field:
Data. The amount and complexity of core resources used to train and improve algorithms.
Total broadband consumption (fixed and mobile): The overall data consumption of a country.
Broadband consumption per capita (fixed and mobile): The amount of data used per Internet user in a country, representing the complexity of different types of data used.
Rules. How to get the data.
Open data participation: The extent to which an economy facilitates the use of and access to public data sources.
Data governance policy: The way in which the state regulates data (personal data, non-personal data, open data, proprietary data, public data, and private data), especially with regard to privacy protection.
Cross-border data flows: The extent to which an economy facilitates and participates in data flows to and from other economies, and the extent to which an economy actively localizes data within its borders.
Capital. Build human, financial, diverse, and digital infrastructure for AI.
Talent: The quality and quantity of existing AI talent.
Investment: Money flowing into AI and emerging technologies.
Diversity: The diversity of AI talent.
The Evolution of the Digital Economy: The Evolution of a Country's Digital Infrastructure, Including Computing Power.
Innovation. Advancements in AI models, technologies, creative data**, and new applications.
Number of patent applications: The number of patent applications for AI technology in each country.
Citations of AI** in the top 10 countries: The total number of citations by authors in each country.
Total AI publications: The total number of publications in the field of AI by country.
Figure 1 Top AI country rankings).
Combining the above variables, HBR derives a new measure, the Top Ranked AI Nations, (TRAIN) to assess AI leadership in 25 countries (as shown in Figure 1), which is not a complete list of countries that play a key role in shaping the global AI industry, as there are data thresholds for the drivers that HBR considers. For example, countries such as Israel or the United Arab Emirates are important players, but they are not included in the report's assessment because they still have little influence.
Leaders in artificial intelligence: the United States and China.
It's no surprise that the U.S. and China are in the top two places on the TRAIN Index, as both countries** are committed to becoming global AI leaders. Jake Sullivan, an adviser to the United States, declared that the United States aims to ensure that it is "as far ahead as possible" in AI cutting-edge technology. At the same time, China aims to become the world's "most important" AI innovation hub, with total AI-related output exceeding 10 trillion yuan by 2030. The White House banned U.S. companies from exporting chip-making equipment in October 2022, a requirement that will now be extended to AI chips, further intensifying competition.
The U.S. leads the way in most key drivers, and the U.S. AI business model is relatively better than that of Chinese competitors. From the perspective of capital factors alone, the top four cities in the world in terms of AI talent, investment, talent diversity, and digital economy development are all in the United States, while Beijing, the number one city in China, ranks eighth. In 2022, venture capital funded 524 AI startups in the U.S., far ahead of all other countries. Over the past decade, U.S. AI companies have attracted 2.2 percent of private investment from China5 times. The private sector is a major driver of AI in the United States. The private sector's share of the largest AI models has soared from 11% in 2010 to 96% in 2021, with 70% of PhDs in AI-related fields employed by the private sector. Fierce competition among U.S. AI companies is likely to continue to drive the U.S. leadership in AI innovation.
In China, ** is playing a greater role in the development of AI. Take advantage of substantial subsidies, support, and policy guidance to channel them into applications such as drug development, genetic research, and biology. China has the largest internet population in the world, and as a result, AI adoption is very fast. For example, China's generative AI tool Wenxin Yiyan reached 1 million users in 19 hours, while ChatGPT took five days to reach the same scale. China has several important advantages that could allow it to challenge the United States in the future:
First, China generates and consumes a lot of data. As a result, China will increasingly have faster-growing data pools that are the hardest to access for AI developers outside of China, a factor that could both hinder and strengthen China's AI leadership;
Second, China has the fastest growing AI research community in the world, with Chinese authors submitting about 2 more to top AI journals than the U.S5 times.
Third, China is a pioneer in AI regulation, and even with the recent White House executive order on AI regulation, China will remain an experienced leader in this area. So far, AI regulation is still in its infancy and does not make up a high percentage of the TRAIN index, but this could change over time.
In addition, China has a number of challenges to contend with: First, strict regulation could weaken its ability to innovate;Second, chip scarcity is a key constraint in the near term;Finally, restrictions on cross-border data flows could hinder China's ability to develop cutting-edge AI models.
The situation of AI leadership in other countries.
Of course, China and the United States are not the only countries leading in AI leadership, and the ranking of TRAIN is not set in stone. With the development of the AI field, there are several other countries to pay attention to, such as the United Kingdom, India, Japan, South Korea, Indonesia, and major countries in the European Union. Some of these countries have faster data pools, while others have easier access to data. In addition, there are demographic factors that will also affect its position in the TRAIN index.
Of all the drivers that are critical to AI leadership right now, changes in the available data pool are likely to have the greatest impact on the rankings of 25 countries in the near and medium term. To understand some of the key changes to note, as shown in Figure 2, HBR mapped the size and momentum of the aggregated data pool on the x-axis to the current train scores of countries on the y-axis.
Figure 2 The size of the aggregate data pool is represented by the size of the circle, and the data availability is indicated by color: red for low data fetching and green for high data fetching).
While capital factors play a powerful role in a country's AI leadership, changes in its data pool are also critical to its ability to move up the train rankings, as Figure 2 shows, with the fastest-growing data pool also having more data access restrictions. This means that companies should be aware of changes in data-related regulations and policies in different parts of the world, and they can decide whether they should move AI development activities to other regions. In addition, policymakers in countries affected by these changes must rethink their own regulations and investment priorities to maintain and strengthen their leadership in AI.
In addition to China and the United States, countries such as India, the United Kingdom, France, Canada, Germany, and Australia have also invested heavily in the AI field. As a result, these countries are expected to make the most significant progress in the medium term. At the same time, the growth of the data pool is also an important factor, with countries such as Indonesia, South Africa, Nigeria, and India having the largest rate of change in total data consumption, so they should be closely watched. Enhancing the availability of data could improve the position of these countries in the future train.
a) India. India has advantages in many ways and is the country with the most potential for development. India is expected to rank first in the world in data consumption by 2028. India has processed more digital payments than any other country in the world and has the world's third-largest AI talent pool. While the country has imposed restrictions on data access, its AI regulatory rules remain volatile. In July, India's Telecom Regulatory Authority of India (TRAI) released a new document calling for the creation of a regulator and a "risk-based framework" to regulate AI in India. The document also proposes to work with countries** and international agencies to advance the "responsible use of AI" globally, which makes India likely to play an important role in this process.
ii) United Kingdom. As for the UK, observing how it competes with EU countries, especially France and Germany, reveals the pros and cons of two very different approaches. The UK's AI industry is supported by a national strategy in the UK, but is committed to a relaxed regulatory approach, aiming to "support innovation" as the sector evolves. In fact, the UK is one of the most innovative AI countries and is home to the likes of DeepMind (which has been acquired by Alphabet), whose research on protein structures could have a groundbreaking impact on areas ranging from drug discovery to food safety. The UK has been working to reconcile its lax regulatory policies with leading AI security, including a Bletchley Declaration to encourage global cooperation on this issue, and the launch of the UK AI Security Institute to conduct security assessments of cutting-edge AI systems.
c) European Union. In contrast, the EU's AI Act is likely to slow down AI development in member states when it is eventually implemented. This will help the UK maintain its current lead over countries such as France and Germany, which have committed to building a common AI ecosystem through a joint statement and the establishment of an "AI Research and Innovation Network". Data pools in France and Germany are growing faster than in the UK, offsetting potential regulatory headwinds, and there has been a heated debate over rethinking the stringency of EU regulations due to lobbying from the industry. Eventually, the regulations adopted a compromise two-tier approach, requiring all but the largest base model to be "transparent." The dismissal and reinstatement of OpenAI's CEO, Sam Altman, could create an opportunity for European AI startups to position Europe as a "more trustworthy AI" company.
4) Japan and South Korea. Both countries have a strong need to develop AI due to their own demographic data and growth priorities, and have invested heavily in this area. Both Japan and South Korea have invested in robots and AI to assist humans in their work. Going forward, both Japan and South Korea face headwinds, as their data pools are not growing at the same rate as some other Asian countries with young populations and growing Internet users.
According to Japan's Ministry of Economy, Trade and Industry, by 2030, Japan will face 78A talent gap of 90,000 software engineers, which in turn will limit their capabilities in deep learning and software development, and slow down their growth in generative AI. In addition, Japan lacks sufficient supercomputing power, which is another bottleneck hindering its AI development. The Japanese venture capital firm SoftBank Group intends to help the country move from "defensive mode" to "offensive mode", but the probability of its success remains to be seen.
South Korea's semiconductor industry and its leadership in AI patents and research can put it at an edge for the time being, however, policymakers and experts are concerned that talent shortages and lack of support will pose huge challenges to South Korea. According to a South Korean lawmaker, South Korea's AI R&D budget has been cut by 43%. From 2013 to 2022, South Korea's cumulative investment in AI was only $5.5 billion, far behind the UK's $18.2 billion. Moreover, compared with the 26 AI-listed companies in Japan, the 6 AI-related listed companies in South Korea are also much inferior.
Summary. Given the potential scope and impact of AI technology, it's no surprise that there is a race for AI leadership across the globe. As global AI power centers emerge and shift, leading countries in the AI field will decide which AI applications are prioritized, which social and economic sectors benefit the most, which data is used to train algorithms, which biases are included, which biases are neutralized, and how to strike a balance between accelerating AI innovation and establishing safeguards. Policy leaders and the business community must pay attention to it, as the geographic location of AI will determine the future of AI and its usefulness to local communities.