Who is leading in artificial intelligence (AI) and machine learning (ML)? How should leadership in AI be evaluated or measured? Which aspects of comparative advantage in AI possess the greatest strategic importance? These questions are critical to address as nations around the world embrace the potential of AI through a range of policy initiatives.
None of these questions yields easy answers. Leadership and comparative advantage in artificial intelligence are difficult concepts to measure. There is no one formula to determine who may be “winning” or will be leading in the long term across various aspects of the field. On some fronts, the United States remains in a relatively favorable position in AI, but its centrality in the ecosystem should not be assumed or taken for granted. The scope and scale of Chinese research in AI are rapidly increasing. Careful evaluation of relative strengths and weaknesses can generate more useful and actionable insights to assess policy choices.
Competitiveness in AI will reflect the dynamism of national innovation ecosystems.Andrew Imbrie, Elsa B. Kania and Lorand Laskai
The United States possesses distinct strengths in top AI talent and research. U.S. comparative advantages reflect its dynamic innovation ecosystem and capabilities in semiconductors. Such advantages can take decades to build and appear to be difficult to buy or quickly duplicate. While the People’s Republic of China (PRC) excels in commercial applications, American prominence in foundational elements and enablers of AI, including hardware, talent, and basic research, are important. Despite considerable progress in AI research in recent years, breakthroughs rarely occur in single moments. The latest advances are the product of decades of refinements to deep learning’s conceptual architecture. Future progress in AI will look less like the space race and instead require dynamic research environments that create and sustain synergies among government, industry, and academia.1
China’s future trajectory in AI remains uncertain. The development of AI in China will depend on the evolution of its overall environment for innovation. The Chinese government is devoting billions to AI through R&D initiatives and government guidance funds, which are stimulating private investments and expenditures by leading companies. These investments may prove effective despite likely inefficiencies in allocation, but also run the risk of introducing new distortions in the market through the surge in funding. The inflated valuations of China’s “AI unicorns” could be a symptom of an “AI bubble.” Looking ahead, the state of AI in China will be hard to disentangle from the broader macroeconomic environment.
This policy brief examines a number of potential strengths for the United States and PRC in AI. Our analysis identifies both areas of U.S. comparative advantage and those where it risks falling behind a rising China.2 Success in AI research, development, and applications will be shaped by the three building blocks of AI: hardware (e.g. AI chips that enable the underlying computing capabilities), the availability of data, and continued advances in algorithms. On the policy and commercial fronts, enablers of AI development include the workforce of AI researchers and engineers, availability of funding for basic and applied research, and private sector investments. Overall, competitiveness in AI will reflect the dynamism of national innovation ecosystems, which we consider in terms of educational opportunities, access to global talent through immigration, and networks of research collaboration.3 The creation of norms and frameworks for governance of AI are equally imperative, while the application of AI to enable a range of military capabilities could affect the future balance of power among nations.4
The state of AI as a field is dynamic and rapidly evolving. In summary, this brief can draw some initial conclusions about the state of play between the United States and China.
The Question of Comparative Advantage in Artificial IntelligenceDownload Full Policy Brief
- K. A. Konrad, “Dynamic Contests and the Discouragement Effect,” Revue d’Économie Politique (2012); C. Harris and J. Vickers, “Racing with Uncertainty,” Review of Economic Studies 54, 1 (1987); I. K. Wang, L. Qian, and M. Lehrer, “From Technology Race to Technology Marathon: A Behavioral Explanation of Technology Advancement,” European Management Journal 35, Issue 2 (April 2017): 187-197.
- This policy brief is not intended to be comprehensive, but rather proposes a framework for assessing relevant data and measures that bear on current debates in AI. We are indebted to the robust research and existing literature in the field. See, e.g., Michael C. Horowitz, Gregory Allen, Elsa Kania, and Paul Scharre, “Strategic Competition in an Era of Artificial Intelligence,” Center for a New American Security, July 2018, 8.
- Deborah J. Jackson, “What is an innovation ecosystem,” National Science Foundation, 1, 2011. On AI in particular, see “AI is a national security priority — here’s how we cultivate it,” The Hill, February 20, 2019, https://thehill.com/opinion/cybersecurity/430765-ai-is-a- national-security-priority-heres-how-we-cultivate-it; “Artificial Intelligence and National Security: The Importance of the AI Ecosystem,” Center for Strategic and International Studies, November 5, 2018, https://www.csis.org/analysis/artificial-intelligence-and-national- security-importance-ai-ecosystem.
- Andrew Imbrie, “Mapping the Terrain: AI Governance and the Future of Power,” Survival (blog), December 17, 2019, https://www.iiss.org/blogs/survival- blog/2019/12/mapping-the-terrain-ai-governance.