Multiple administrations in Seoul and Washington have recognized the importance of cooperation in emerging technologies, particularly in artificial intelligence (AI). The 2021 joint statement, for instance, commits both nations “to collaboratively develop a future-oriented partnership.” This year’s follow-up joint statement further underscores the significance of “strengthening public and private cooperation” by leveraging AI talents. In essence, fostering AI bilateral cooperation aligns with the national interests of both South Korea and the United States. Even within its trilateral context, the urgency of this need is evident in the latest Camp David summit’s call to “maintain focus on building robust cooperation in the economic security and technology spheres.”
However, the call for AI cooperation still lacks clarity, primarily due to the lack of concrete discussions about the necessary steps for making this collaboration a reality. Facilitating cooperation that aligns with national interests requires a mutual approach, as defined by two key aspects according to the international relations scholar Robert Keohane. First and foremost, collaboration relies on a conditional framework similar to quid pro quo dynamics. A reciprocal action follows prior actions taken by the partner and contributes to their shared goals. Following this principle, there's a notion of equivalency, meaning that any assistance received should ideally be met with proportionate gestures or resources.
Exploring AI cooperation between South Korea and the United States raises a fundamental question: Can a reciprocal framework be implemented? Within the context of contingency, the feasibility of applying the principle of “an eye for an eye, a tooth for a tooth” seems uncertain. This challenge is particularly evident in the landscape of private enterprises in both countries, frequently caught in competitive dynamics. A notable example of this competitive spirit is Naver. This major Korean IT company intends to advance a customized AI model to address non-English-speaking nations’ distinct linguistic and politically nuanced needs.
Furthermore, in the South Korean context, six prominent companies have either introduced or are developing their own generative AI models. These efforts are driven by a shared goal to provide personalized and finely-tuned services, particularly for the domestic user base. The ultimate objective of these initiatives is to enhance user experiences by delivering services tailored to the nuances of the local landscape.
The recent conflict stemming from the Inflation Reduction Act, which revolves around consumer subsidies for electric vehicle manufacturers, vividly illustrates the power imbalance. This case highlights the breach of the reciprocity principle. Although South Korea succeeded in having its manufacturers included in the list of eligible recipients for these subsidies, a substantial portion of its domestic subsidy fund ultimately benefits American manufacturers. This situation underscores a violation of the contingent facet of reciprocal cooperation.
For instance, this discrepancy is evident when purchasing a Tesla Model 3 in South Korea. Consumers receive approximately 4 percent of the total vehicle cost as a tax credit. In stark contrast, the principle of reciprocity weakens when attempting to acquire an equivalent electric vehicle, such as the Hyundai Ioniq 6, within American borders. In this context, the absence of available tax credits for purchasing purposes becomes apparent, with a limited 15 percent tax credit reserved solely for leasing agreements. This comparison highlights the inequitable nature of reciprocity in consumer incentives for electric vehicles.
In the broader context of national AI capacity, a discernible disparity exists between the United States and South Korea, as highlighted by data from the Emerging Technology Observatory’s Country Activity Tracker dataset. Notably, the United States stands out as the foremost producer of internationally collaborated articles, contributing to nearly half of all AI-related scientific publications. Interestingly, U.S. researchers collaborate extensively with their Chinese counterparts, co-authoring 32 percent of these articles, while collaboration with South Korean researchers remains relatively modest at 4 percent.
Conversely, South Korean researchers actively collaborate with U.S.-affiliated researchers, accounting for 44 percent of their joint publications. In terms of direct investment in AI companies, South Korea invests $1.7 billion in the U.S. sector, while American investors direct $2 billion. The former accounts for 30.5 percent of Korea’s total disclosed AI investment inflow, whereas the latter represents 0.5 percent of the American total. This analysis accentuates the substantial gap in AI research and investment levels between the United States and South Korea. It also underscores South Korea’s significant reliance on American investment and academic capacity, while the United States may consider South Korea as one of several potential collaborators.
The series of analyses underscores the impracticality of achieving genuine reciprocal AI cooperation between the two countries. This leads to another pertinent question: What should be the guiding principle for such collaboration? Unlike traditional reciprocity, diffuse reciprocity emphasizes moral values like mutual trust and good faith instead of rigid equivalence in exchanges. This innovative approach promotes cooperative endeavors that advance the collective good, transcending the limitations of direct quid pro quo arrangements.
Since they share democratic values, Seoul and Washington must explore collaborative avenues that amplify individual efforts and promote the broader welfare. Diffuse reciprocity emerges as a promising framework to shape the future trajectory of AI and emerging technology collaboration between Seoul and Washington.
Both nations possess the capacity to complement areas where the other may exhibit vulnerabilities or require further advancement. According to OECD.AI data, the United States has an extensive research and development infrastructure, marked by its wide-ranging international research collaboration and significant investments from the public and private sectors. This fusion of academic and industrial strength positions the United States as a formidable leader in technological innovation. On the other hand, South Korea exhibits a resilient AI workforce and the potential to cultivate a conducive research environment, highlighted by its transformation into a hub for AI talent inflow. As such, the convergence of South Korea’s skilled workforce and America’s robust research infrastructure promises synergistic outcomes.
In fostering research networks and nurturing cooperative relationships with other nations, Seoul and Washington find common ground with key partners, including Canada, India, the United Kingdom, Germany, Japan, and France. These shared allies not only prioritize democratic values in their foreign policies but also exhibit a strong commitment to scientific advancement. Leveraging this shared foundation, the diffusion of cooperative efforts between Seoul and Washington can potentially radiate benefits across these interconnected networks.
By infusing their cooperative endeavors with a sense of shared purpose and mutual trust, both nations can pursue ambitious goals that transcend the constraints of traditional reciprocal arrangements. At the same time, this collaborative synergy arises from the inherent power imbalance between the two nations, with each contributing unique strengths to address their respective shortcomings. In doing so, they will strengthen their bilateral partnership and contribute to advancing democratic values and technological progress on a global scale.
Sanghyun Han is a Ph.D. student in International Affairs, Science, and Technology at Georgia Tech’s Sam Nunn School of International Affairs. He has previously worked with research institutions in both Washington D.C. and Seoul, such as the Atlantic Council, the National Bureau of Asian Research, the Institute of Foreign Affairs and National Security, and the East Asia Foundation.
This article was prepared as part of the CSIS Korea Chair’s 3.0 Program, where the author is part of the inaugural cohort. The author would like to thank Victor Cha, Evan Ramstad, and other Young Leaders for their insights.