[Special Report on the New Administration's Foreign Policy Recommendations] ③ New Administration's Technology Diplomacy Strategy Centered on Artificial Intelligence Technology

I. Background

As of 2025, the AI competition is unfolding on all fronts, centered on the US and China's race for leadership in AI models and technological infrastructure, the EU's strategy to lead in regulation, the AI development strategies of the UK, Japan, South Korea, and the Middle East, and the pursuit of regulations on AI weaponry and global AI safety governance. It is no exaggeration to say that AI has become the focal point of the global order restructuring, extending beyond technological levels to encompass economy, security, and norms. Securing and utilizing AI technology has become essential for a nation to achieve economic prosperity, enhance national security, and lead international norms in the 21st century. Nations are striving to increase investment and nurture talent to enhance their AI technological innovation capabilities and utilization, while simultaneously crafting technology diplomacy strategies to expand their influence in AI technology development and application.

In May 2025, during the "Winning the AI Race" hearing held by the U.S. Senate, Sam Altman and other industry leaders unanimously pointed out that AI is a technology that will bring about greater changes than the internet, and that while the U.S. currently holds an advantage, the gap with China, a challenger, is not that large. They particularly emphasized that the US and China are competing not just in technological development but also in disseminating technology worldwide, and that for the US to maintain its leading position, it must lead international cooperation alongside continuous investment in comprehensive areas, including infrastructure and energy.

Early this year, China's startup's generative AI model 'DeepSeek-R1' caused a significant stir as a low-cost, high-efficiency model, drawing attention as an alternative to US-led AI development. Bolstered by DeepSeek, the Chinese government proposed the 'AI+' strategy in the 2025 Two Sessions (两会), designating artificial intelligence (AI) as a core element of its national strategy. The Two Sessions mentioned the concept of 'Embodied Intelligence (具身智能)' and emphasized the promotion of AI application industries centered on manufacturing, such as humanoid robots, connected cars, and AI smartphones. It also stated its intention to foster innovation by applying AI across various industrial sectors beyond manufacturing, including finance, transportation, public services, healthcare, education, and welfare. China is building data centers and computing infrastructure through the 'East Data West Computing (东数西算)' project to process data from eastern regions in western regions. Concurrently, it has been promoting the 'AI Global Governance Initiative' to form a global AI governance framework and standard norms based on broad consensus and respect for the policies and practices of various countries, and to support the establishment of a global AI governance body under the UN. China leads in AI paper and patent submissions, and its AI models are spreading to various regions including the Middle East, Latin America, and Southeast Asia. According to the 2025 Stanford Index, the US developed 40 notable AI models, significantly ahead of China (15) and Europe (3). However, while quantitative superiority is maintained, the performance gap with Chinese models is rapidly narrowing from double digits in 2023 to an almost equivalent level today. As China's AI challenge begins to yield partial results, a notable shift is occurring, creating cracks in the US's previous overwhelming dominance, and whether this trend will continue is a key point to watch.

The European Union (EU), after finalizing its 'AI Act' in late 2024, announced a public and private AI investment plan of 200 billion euros at the 'AI Action Summit' held in Paris, France, in February 2025, striving to lead global norms for 'trustworthy AI.' In April 2025, Singapore hosted an AI Safety Summit attended by major AI companies such as OpenAI, Google DeepMind, Anthropic, and Meta, as well as leading academic institutions and researchers from MIT, Stanford, Tsinghua University, and the Chinese Academy of Sciences. Singapore emphasized its friendly relations with both the US and China, presenting a vision to contribute to the establishment of global AI governance by promoting international cooperation in AI safety and strengthening its role as a neutral mediator in this field. Saudi Arabia is attracting attention by promoting a structural transformation from a resource-dependent economy to a technology-centric nation, establishing the AI company 'Humain' with funds from its Public Investment Fund (PIF), and building Arabic-based AI models and large-scale data centers. Many countries, including the EU, Singapore, Saudi Arabia, Japan, South Korea, Australia, and Canada, are seeking AI strategies to establish and solidify their positions within the landscape of global AI technological development led by the US and China. Last year, a UN General Assembly resolution saw 164 countries agree that the risks of autonomous weapons urgently need to be addressed by the international community. As a follow-up, discussions on regulating AI-based 'killer robots' took place at the UN in May 2025, but a binding agreement was not reached due to differing stances among the US, China, and Russia. Various international organizations, including the UN, OECD, and UNESCO, as well as the G7, individual countries, and corporations, are proposing various measures to ensure AI safety, but these have not yet led to substantial outcomes.

The AI policies of the United States, which is currently leading AI development, significantly influence the direction of global AI deployment. While the previous Biden administration sought to balance AI innovation with safety regulations, the Trump administration's second term, upon taking office, sent a message prioritizing innovation over AI safety regulations by immediately repealing the Biden administration's 'Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence.' It subsequently issued the 'Executive Order on Removing Barriers to American Artificial Intelligence Leadership' and directed the formulation and submission of an AI implementation plan to the President within 180 days to strengthen America's global AI leadership, competitiveness, and national security. Furthermore, in collaboration with OpenAI (AI research and development), Oracle (data management and cloud infrastructure provision), and SoftBank (funding), it announced the large-scale 'Stargate Project,' investing $500 billion over the next four years to build a US-centric AI ecosystem and enhance the competitiveness of the US AI industry in response to China's AI technological advancements, signaling significant investment in the AI sector.

In March 2025, the U.S. Department of Commerce's Bureau of Industry and Security (BIS) announced export controls on 80 companies, including 53 Chinese firms, signaling that export controls on advanced technologies to China would continue under the second Trump administration. BIS cited China's military modernization and support for the development of artificial intelligence and quantum computing technologies as reasons for the export controls. The Biden administration pursued a dual strategy of controlling exports to China (protection), supporting domestic advanced manufacturing capabilities through the CHIPS Act (promotion), and forming technology alliances with the EU, Japan, Taiwan, and South Korea (partnership). This so-called '3P' policy was implemented as a package, effectively curbing China's technological rise. However, the second Trump administration, advocating for 'America First,' is employing a different approach to supporting advanced manufacturing and technology alliances, and it remains to be seen how this will affect the effectiveness of curbing China. The Trump administration argues that imposing tariffs to drive domestic advanced manufacturing investment is more effective than providing subsidies to foreign companies and seeks to achieve desired outcomes through pressure rather than cooperation with allies. Semiconductors were mentioned alongside copper and pharmaceuticals as items not subject to mutual tariffs, but instead, item-specific tariffs were announced, drawing attention to the future direction of tariffs related to semiconductors and artificial intelligence. Recently, the Trump administration imposed a 32% tariff on Taiwanese products, but temporarily exempted semiconductor products, reflecting the importance of Taiwan's semiconductor industry in the US market. Amidst the acceleration of global supply chain restructuring and the decline of the free trade order due to the Trump administration's export controls and tariff pressures, nations face the challenge of strengthening their advanced technological capabilities, restructuring international cooperation frameworks for this purpose, and pursuing their prosperity and security through delicate technology diplomacy.

II. Recommendations

South Korea has developed and pursued an AI strategy based on its strengths in AI technology and semiconductor infrastructure, establishing the AI Basic Act and hosting the 'AI Seoul Summit' in 2024 to propose global AI governance centered on innovation, safety, and inclusivity. However, the measures taken thus far are insufficient to navigate the AI era. In a situation where AI is becoming a focal point for prosperity, security, and values, more ambitious policies are required to enhance South Korea's AI capabilities, utilization levels, and global influence.

First, investment in AI infrastructure and talent, which form the foundation of technological diplomacy, must be significantly expanded and supplemented. South Korea developed its semiconductor technological innovation capabilities through bold investments in the early 1980s, which have become our most important diplomatic asset today. It is clear that the AI technology we secure will become a crucial diplomatic card in the future. Investment for this purpose must be made now. South Korea's total AI investment in 2024, including government and private sectors, is estimated at approximately 3.7 trillion KRW, which is significantly insufficient compared to the US ($190 trillion KRW) and China ($52 trillion KRW), even considering the difference in economic scale. The government announced an ambitious plan last year to invest a total of 65 trillion KRW by 2027, but the investment performance to date remains very modest compared to the plan and to other countries. A prominent example of insufficient AI investment is evident in the number of latest AI chips, NVIDIA's H100, secured. Securing cutting-edge chips is crucial for continuous AI technological development; currently, South Korea has secured about 2,000 units, which is far behind the millions in the US and the estimated 100,000 in China. A concrete technological diplomacy strategy must be established for securing chips, along with securing financial resources. Regarding domestic AI talent, not only is there an absolute shortage, but it is also revealed that 40% of domestic AI talent has emigrated, which will negatively impact South Korea's AI development in the long term, necessitating urgent countermeasures. Furthermore, institutional conditions such as energy acquisition and personal data protection, which are crucial for the diffusion of AI technology, must be developed concurrently.

Technological diplomacy encompasses two aspects: utilizing diplomacy for technological advancement and utilizing technology for diplomatic communication or outcomes. We question whether diplomacy is important for the advancement and expanded utilization of AI technology. Primarily, domestic investment, talent development, and various support policies for technological advancement are important. However, it is evident that technological development does not occur in a vacuum or within a limited domestic space. R&D activities themselves often occur through international exchange, requiring the import of core technologies, necessary materials, and equipment, and also involving the export of developed technological products. During periods when the liberal world economic order functioned smoothly, cross-border technology exchange occurred overwhelmingly in terms of cost and efficiency. However, in times of instability in the liberal world political and economic order, characterized by export controls and tariff increases, diplomatic and security considerations outweigh the logic of cost and efficiency. Therefore, when structuring international cooperation frameworks for AI technology development and deciding which countries to cooperate more closely with, it is inevitable to consider both the internal technological situation and the diplomatic-security situation, making technological diplomacy crucial. This section presents specific proposals for the new administration to pursue from the perspective of AI technology diplomacy.

First, let's examine the internal technological situation through AI semiconductors. South Korea's memory and system semiconductor sectors depend on the US for major manufacturing equipment and IP. The US market accounts for less than 10% of South Korea's semiconductor exports. The latest chip, H100, used for running AI models, is designed by NVIDIA in the US and manufactured by TSMC in Taiwan. On the other hand, South Korea imports a significant portion of key raw materials such as silicon, germanium, and tungsten from China, and over half of South Korea's semiconductor exports go to China (including Hong Kong). Samsung and SK Hynix have manufacturing facilities in China. If cooperation with the US is halted, and South Korea is unable to receive semiconductor equipment and the latest chips, which are powerful cards held by the US, South Korea's semiconductor production and AI technology development will immediately cease. Conversely, a reduction in cooperation with China will disrupt South Korea's semiconductor production and exports. To date, China has needed South Korea's technology and semiconductor imports, thus not exerting significant pressure on South Korea, but the fact that South Korea's share in China's semiconductor imports is gradually decreasing is problematic. China's most threatening card is catching up to South Korea's semiconductor manufacturing technology. Over the past few years, US export controls have slowed down China's semiconductor manufacturing technology innovation, buying time for South Korean companies. Considering these circumstances, the most important partner for South Korea's technology diplomacy will inevitably be the United States for the time being.

In terms of external technological factors, the relationship between the liberal world economic order and South Korea's security and prosperity is particularly noteworthy. South Korea's economic growth, which began in earnest in the 1970s, was built upon the foundation of the ROK-US alliance and the free trade order. South Korea's technological innovation capabilities were also developed within the US-led open global innovation system. Currently, with the US-China competition and the US's export controls and tariff increases, the liberal world economic order is declining, and the globalization and innovation systems are becoming bloc-based, leading to side effects such as price increases, economic recession, and rising R&D costs. The South Korean economy is at the forefront of experiencing the shocks from the decline of the liberal economic order and the bloc formation of global supply chains and innovation systems, and must strive to restore and reorganize the rule-based liberal economic order for shared prosperity in any form. In the so-called symbiotic liberal world economic order, which we build together for shared prosperity, the United States is a more important partner than any other country for South Korea in terms of the complementary aspects of both economies. On one hand, we must respond to US pressure, and on the other hand, we must seek the reorganization of the liberal world economy by cooperating with liberal democratic countries, including the United States.

The United States is leading global AI technological development, possesses technologies that we lack, and is a partner with whom we must cooperate for the reorganization of a symbiotic liberal world economic order; thus, it is the most important target for South Korea's technology diplomacy. The ROK and the US launched the Next Generation Critical and Emerging Technologies Dialogue and the Korea-US AI Working Group, setting key cooperation areas such as machine learning and AI development, international AI standardization, research cooperation, policy interoperability, and international AI governance, but substantive cooperation has not progressed. In 2024, the US hosted the International Network of AI Safety Institutes meeting in San Francisco, where both countries are cooperating on evaluating the safety of AI models and developing international standards. The new administration must support bilateral AI cooperation among various actors, including government agencies, research institutes, universities, and businesses, concerning AI technology development and safety.

In an unstable international order, expanding advanced technology cooperation with the United States, an ally and holder of core technologies, is crucial. Particularly, as a counter-benefit to pressure regarding defense cost-sharing and tariffs, we must proactively present and implement new technology cooperation agendas to the US, which may have weaker incentives for cooperation than we do. Currently, R&D budgets are being reduced under the US government's efficiency drive. Amidst this, cooperation should be expanded in areas that the Trump administration actively shows interest in and areas that can be utilized for containing China. A system should be established to lead and monitor ROK-US advanced technology cooperation. While past ROK-US advanced technology cooperation, focused on the objective of science and technology itself, allowed for dispersed cooperation with no significant issues, current strategic cooperation, conducted from the perspective of overall national security or foreign policy strategy, requires a focal point for technology diplomacy to comprehensively monitor and coordinate such strategic cooperation.

In the context of US-China competition, China's AI technological advancements necessitate a deeper consideration of how to conduct technological diplomacy with China, raising the need for more complex and multi-layered diplomatic capabilities. While the space for cooperation with China is indeed shrinking due to strengthened cooperation with the US, it is necessary to maintain technological diplomacy channels with China and sow the seeds for future concrete cooperation. It is essential to restore and manage diplomatic engagement with China, allowing networks and cooperation between universities, research institutes, and businesses to be maintained and developed, particularly in strategically less sensitive science and technology sectors, thereby opening the door for future collaboration.

With the acceleration of AI technology development, demands for ensuring safety and protecting personal information are increasing. South Korea, following the European Union, became the second country in the world to enact the 'Framework Act on the Development and Dissemination of Artificial Intelligence and the Promotion of Trustworthy Artificial Intelligence' (commonly known as the AI Basic Act) in 2024. This act aims to promote the development and utilization of AI while simultaneously pursuing ethics, safety, and the establishment of a trustworthy ecosystem. Balancing AI promotion and regulation, its key provisions include mandatory establishment and renewal of national AI strategies, risk-based AI classification and government response guidelines, expanded opening of public sector data, mandatory adherence to AI ethical principles, and the establishment of evaluation criteria. International organizations such as the UN, OECD, UNESCO, and G7, as well as countries including the US, China, the UK, the Netherlands, Singapore, and France, have proposed global AI norms and governance frameworks. South Korea is actively participating in GPAI (Global Partnership on AI), OECD AI Principles, G7 AI Process, and UN AI governance discussions. In 2024, it hosted the 'AI Seoul Summit,' proposing safety, innovation, and inclusivity as AI governance values through the Seoul Declaration. Having taken the initial step, it is crucial to continuously participate in and lead global AI governance discussions that balance AI development and trust while reducing disparities.

Amidst the expansion of influence by giant AI companies and major powers in the US and China, many countries are interested in developing Sovereign AI based on Large Language Models (LLMs) that manage their own data and infrastructure, and reflect their local languages, cultures, and values. As AI technological innovation is increasingly led by a few big tech companies, countries lacking their own independent platforms or LLMs are beginning to recognize the threats posed by AI outputs that do not reflect their national identity or interests. Consequently, Sovereign AI strategies for building independent national AI capabilities are being actively discussed, with a high demand for strengthening supercomputing capabilities and developing proprietary AI models. Attempts are underway, such as France's 'Mistral AI' and 'LeChat,' Finland's 'Silo AI,' the UK's 'BritGPT' designed to focus on British culture and history, and the development of Japanese-language specialized LLMs in cooperation with NVIDIA. Discussions on developing and expanding Korean-style Sovereign AI are also actively underway domestically. Naver has developed 'HyperCLOVA X,' based on Korean language and reflecting Korean socio-cultural contexts, and is pursuing cooperation in developing sovereign cloud and sovereign AI optimized for the Middle East and Southeast Asia based on this. To solidify South Korea's influence in the global AI ecosystem, successful development and expansion of Sovereign AI are crucial, requiring active government support. Furthermore, we must expand bilateral/multilateral cooperation in AI with the Global South and develop and present packages to support the comprehensive AI transition of the Global South. Currently, cooperation with the Global South, particularly ASEAN countries, is actively progressing in the cybersecurity field. In November 2024, South Korea and ASEAN were elevated to the status of Comprehensive Strategic Partners, and cooperation is underway in various fields including cybersecurity, defense, supply chains, and digital transformation. Building on this, cooperation in the AI sector should also proceed, expanding technological cooperation with developing countries through initiatives like the Korea-ASEAN Digital Flagship Project, and extending cooperation beyond ASEAN to the Middle East, Africa, and South America.

The starting point for South Korea's technology diplomacy lies in contemplating what kind of nation and world order we aim to build through artificial intelligence. This will inevitably differ from the paths pursued by the US, led by big tech, or China, led by an authoritarian state. If not the American or Chinese model, then discussions must earnestly take place to define the Korean model, leading to social consensus. This is a core aspect of South Korea's future identity, encompassing not only the development of Sovereign AI but also the content and values it will embody. Despite numerous warnings about the challenges and risks posed by AI, the nature of the threat remains unclear. Lu Xun described a situation where hostile forces surround us, but a clear enemy cannot be identified, friendly and enemy lines are blurred, and a distinct front cannot be formed, as 'Mu Wu Zhi Zhen' (無物之陣, a formation without objects). AI, an overwhelming, somewhat alluring, and simultaneously threatening entity whose true nature cannot yet be fully grasped, is driving us towards such a state. What is clear so far is that the decline of the liberal world economic order, which has supported South Korea's security and economic growth, is progressing, and the establishment of effective global AI governance to counter AI threats is also uncertain. The reorganization into a rule-based liberal world order for co-prosperity, which is essential for South Korea's continued prosperity and security, can be a shared aspiration for many countries amidst an unstable and uncertain global order, not just South Korea's wish. Amidst the ongoing US-China hegemony competition, we must pursue technology diplomacy that, on one hand, strives to lead in AI technology development and utilization while focusing on our security and prosperity for reorganization into a co-prosperous liberal world order, and on the other hand, identifies and collectively responds to the challenges and risks posed by US-China competition and AI by consolidating efforts and forging alliances.

■ Bae Young-ja_Professor, Department of Political Science and International Relations, Konkuk University.

■ Managed and Edited by:Song Chae-rin, EAI Research Fellow

    Inquiries and Editing: 02 2277 1683 (ext. 211) | crsong@eai.or.kr