Husanjot Chahal, Ngor Luong, Sara Abdulla and Margarita Konaev
Executive Summary
The Quadrilateral Security Dialogue, commonly known as the Quad, is a semi-formal but strategically significant grouping of four countries—the United States, Australia, India, and Japan. Cooperation on critical and emerging technologies is a key element of the Quad’s agenda, and all four nations have a particular interest in strengthening cooperation on responsible development of artificial intelligence (AI).1 Their desire to collaborate stems not only from recognizing AI’s transformative economic, societal, and national security potential, but also the importance of ensuring that technological innovation is shaped by their shared democratic values and respect for human rights.2 The Quad could offer an alternative to China’s techno-authoritarian model of technology development and use, setting the standard for a multilateral approach to countering the malicious use of AI for surveillance, censorship, and misinformation.3 The group, however, faces non-negligible barriers to effective technology cooperation, including different approaches to data governance, varying economic and technological capabilities, and divergent geopolitical priorities.
This report assesses the state of AI collaboration among the Quad members, focusing specifically on trends in joint AI-related research publications and investment flows into AI companies over the past decade. It also evaluates the AI-related research and investment ties between each of the Quad countries and China during this period. Our main findings are as follows:
AI Research:All four Quad countries are among the top 10 AI research producers in the world, both by number of research publications and citations. Researchers in the Quad countries collectively generated nearly 650,000 AI-related research papers between 2010 and 2020, more than the total authored by European Union and Association of Southeast Asian Nations (ASEAN) researchers combined.
Each Quad country has AI-related research strengths that could be leveraged for joint research opportunities. Japan stands out in simulation and human-computer interaction (HCI), India in data mining and data science, Australia in linguistics and theoretical computer science, and the United States in machine learning and natural language processing.
While the United States collaborates extensively with Australia, India, and Japan on AI-related research, the latter three Indo-Pacific states collaborate little with one another. The United States is the leading research partner for Australia, India, and Japan, and AI researchers from these three Indo-Pacific countries collaborated with U.S. peers on at least 19 percent of their internationally co-authored research papers. In contrast, collaboration rates between Australian, Indian, and Japanese researchers never exceeded 4 percent of each country’s respective internationally co-authored AI research output.
China is the top research partner for the United States and is the second-leading partner for the rest of the Quad members. Not only does research cooperation between the United States and China outweigh U.S. research collaboration with the rest of its Quad partners taken together, but Australia, India, and Japan each have more research partnerships with China than they do with one another.
AI Investments:Between 2010 and 2021, the majority of investment transactions in AI companies located in the United States, Australia, India, and Japan included domestic investors. The United States, however, is the largest foreign investor in Australian, Indian, and Japanese AI companies, both in terms of the number of investment transactions and overall transaction value.
While the United States has robust AI investment ties with Australia, India, and Japan, there is relatively little investment activity between the latter three Indo-Pacific countries. Although there was more AI investment activity between Japan and India than between India and Australia or Australia and Japan, it was largely one-sided, with Japanese investors targeting Indian AI companies while Indian investors seem more reluctant to pursue opportunities in Japan.
There is far more AI investment activity between the United States and China than the United States and each of the remaining Quad countries. While the number of Chinese investment transactions in the U.S.-based AI companies has declined since their peak in 2017, U.S. investments in China’s AI companies increased in 2021.
Australia, India, and Japan each have more AI investment activity with China than they do with one another. Similar to the trends observed in the Quad countries’ AI research collaboration, AI investment activity between China and each of three Indo-Pacific members of the Quad exceeds the limited investment flows between these three countries, both in terms of the number of investment transactions and overall transaction value.
The Quad offers a forum to build trust, identify opportunities for joint research ventures, and gather AI entrepreneurs, investors, and strategic industry partners to increase and diversify technology collaboration. But the prospects for its success depend largely on building stronger ties among U.S. allies beyond their bilateral linkages to the United States. Patterns of AI-related research collaboration and investment across the Quad highlight that the group’s three Indo-Pacific members are less closely intertwined with one another than they are with the United States. Moreover, each of the Quad states has varying but fruitful AI research and investment relationships with China. To capitalize on the Quad’s potential, fuel innovation, and decrease dependency on Chinese technology and markets, the group needs to strengthen and expand AI research collaboration and investment among Australia, India, and Japan.
No comments:
Post a Comment