[New Year Special Commentary Series] ⑨ Global Security Landscape 2025: The Military Use of Artificial Intelligence and the Cult of the Offensive

I. Outlook for the Global Military Environment in the Trump 2.0 Era

In five days, President Trump will return to the White House. From the Prime Ministers or Presidents of Canada, Argentina, Hungary, and Italy to the CEOs of Apple, OpenAI, Meta, Amazon, and SoftBank Group of Japan, leaders are rushing to visit Mar-a-Lago, where the presidential transition team is being formed. It is analyzed that various national and corporate leaders are desperately trying to avoid the wrath of Trump, who has announced a universal tariff of over 10% (Lee Jae-rim 2025). Meanwhile, South Korea faces an unprecedented situation where, following the impeachment motion against President Yoon Suk-yeol passed on December 14, the Prime Minister and Acting President Han Duck-soo's duties were suspended on December 27, leaving Deputy Prime Minister and Minister of Economy and Finance Choi Sang-mok to act as the acting acting president. If, as projected by major U.S. think tanks, U.S. policies on key issues concerning Korea—namely, the presence of U.S. Forces Korea, tariffs, and the semiconductor law—are decided within 100 hours of Trump's inauguration (Park Sung-min 2024), South Korea, lacking a representative leadership to visit Mar-a-Lago, must be considered to be starting 2025 from a disadvantageous position compared to other nations.

However, it can also be argued that structural variables operating at a deeper level of the global security order, independent of leadership variables such as a second Trump term or South Korea's impeachment crisis, are in fact more significant (Waltz 1979). As former British Prime Minister Clement Attlee stated, “Revolutions do not change geography, and revolutions do not change geographical needs.” Regardless of Trump's personal idiosyncrasies, the fundamental nature of U.S. strategic interests remains unchanged, and South Korea's national interests, to a certain extent, exist separately from domestic political turmoil. Particularly, considering the transactional approach of Trumpism, which prioritizes a clear transactional approach over rule- or value-based diplomacy (Jeon Jae-seong 2025), structural changes in the actual reality, and the strategic interests of each nation formed within that context, may be more important than individual leadership variables. In preparation for the period when South Korea overcomes its domestic political turmoil and resumes normal foreign and security policy operations, an analysis of these structural variables that will shape the global security environment is necessary.

This paper examines the direction of change in the global military order in 2025 from the perspective of military technological change, rather than great power strategy, leadership characteristics, or domestic political factors. Specifically, it reviews the changes that the military use of Artificial Intelligence (AI) is currently bringing about in terms of the offense-defense balance, examining autonomous weapon systems, cybersecurity, and the nuclear-AI nexus. It also analyzes the reasons for the recent emergence of a 21st-century version of the 'cult of the offensive' phenomenon within the policy communities of the United States and China. Furthermore, it argues that ultimately, the perception of military technological change, rather than the change itself, becomes a crucial factor determining the nature of the future global security environment, and proposes directions for South Korea's security policy.

II. Military Technological Change: Reality vs. Perception of AI's Military Use

Among the structural variables that influence changes in the military security order, distinct from leadership or domestic political variables, the 'offense-defense balance' is the most critical. This concerns the question of which side, offense or defense, is more effective when investing the same amount of resources. Robert Jervis emphasized that the severity of the security dilemma, which critically impacts the ease of international cooperation, is determined by this offense-defense balance (Jervis 1978, 187-199). In various international crisis situations, offensive advantage has tended to increase the risk of escalation, while defensive advantage has tended to enhance strategic stability. For this reason, the offense-defense balance has broad implications for security dynamics, including the likelihood of war, alliance politics, and arms races (Kim Yang-gyu 2024a).

In the medium to long term, the most noteworthy technology influencing changes in the future offense-defense balance is AI and its military applications. AI functions as an 'enabler' that drives changes in other technologies and a 'force multiplier' that enhances the effectiveness of existing technologies, possessing characteristics of a foundation for technological development and a meta-technology (Horowitz 2018). This is precisely why AI is at the core of the U.S. de-risking strategy, described as a 'small yard, high fence,' and the U.S.-China high-tech competition (Son Yeol et al. 2023; Kim Yang-gyu 2024a; Bae Young-ja 2025). Recent U.S. strategic documents are predicated on the expectation that AI technology will be a key means for maintaining U.S. hegemony and its overwhelming advantage over China (Kim Yang-gyu 2024b; Jacobsen and Liebetrau 2023).

AI is predicted to be a 'force enhancer' that analyzes vast amounts of data to make rapid and accurate decisions, increasing efficiency in military operations. It is expected to bring about innovative changes, particularly in areas such as battlefield information processing, target detection, and enhancing the speed of response to enemy tactics. Furthermore, AI plays a crucial role in strengthening multi-domain operational capabilities and enabling integrated operations across land, sea, air, space, and cyber domains, as exemplified by the U.S. concept of 'integrated deterrence' and China's 'intelligentized warfare' (Son Yeol et al. 2023; Kim Yang-gyu 2024b; Bae Young-ja 2025).

1. Objective Reality of the Global Security Environment: AI-Cyber, AI-Autonomous Weapon Systems, Nuclear-AI Nexus

What changes will the military use of AI bring to the offense-defense balance? To answer this, it is necessary to examine the recent changes occurring with the integration of AI into existing forces in terms of cyber attack/defense capabilities, autonomous weapon systems, and the nuclear-AI nexus (Kim Yang-gyu 2024b; Jeon Jae-seong 2024).

First, in the realm of cybersecurity, AI is enhancing both offensive and defensive capabilities (Jacobsen and Liebetrau 2023). However, many current discussions on this issue, particularly in government documents, more readily assert that AI provides an offensive advantage in the cybersecurity balance. This is partly due to the inherent nature of the technology, which blurs the lines between peacetime and wartime cyber attacks, and the practical difficulties in developing capabilities to detect and defend against such attacks (Jeon Jae-seong 2024). For instance, in responding to malware, there is a technical challenge in detecting it when its purpose and target are ambiguous. Exploiting this, sophisticated AI-driven malware attack programs like Deeplocker have been developed. This program operates by hiding its intent until it reaches a specific victim, and then, upon identifying the target through indicators such as facial or voice recognition by the AI model, it deploys the malware, making detection and response considerably difficult.

To counter AI-enhanced cyber attack programs, defensive programs integrated with AI capabilities are equally necessary. Currently, cybersecurity companies are developing both AI-integrated offensive and defensive programs. This is because defensive programs cannot be developed without understanding how an adversary might utilize cyber attack programs. Thus, within the single actor of a cybersecurity program developer, a continuous contest between offense and defense is taking place. In particular, the paradox where strengthening AI integration to enhance cyber defense capabilities increases system interconnectivity and expands the 'attack surface' that adversaries can target, thereby deepening the vulnerability of the entire defense system, makes it extremely difficult to definitively discuss which side, offense or defense, holds the advantage in cybersecurity.

Second, similar to the cybersecurity domain, AI's application in autonomous weapon systems is unlikely to unilaterally create an advantage for either offense or defense. The primary reason autonomous weapon systems are receiving significant attention is their potential to drastically reduce the costs of warfare. As demonstrated by drone and killer robot technologies, autonomous weapon systems are mass-producible, highly scalable and proliferable, and allow a single operator to manage multiple weapon systems simultaneously. Consequently, the scale of damage that can be inflicted on an adversary is determined by the number of weapons themselves, not the number of operators. Each robot can operate independently while maintaining efficiency, allowing for the elimination of specific targets with minimal collateral damage (Jeon Jae-seong 2024). Furthermore, by minimizing the involvement of human soldiers required for the success of military operations, concerns about increasing casualties due to prolonged warfare or resulting domestic public backlash can be alleviated. Some research argues that the reduction in warfare costs brought about by AI-enhanced autonomous weapon systems represents a true military innovation (Schneider and Macdonald 2024). The reduction in the cost of waging war can be seen as strengthening the offensive advantage in the offense-defense balance.

However, AI-enhanced autonomous weapon systems also exhibit characteristics that provide a defensive advantage. For example, considering a conflict between the U.S. and China centered around the Taiwan Strait, recent technological advancements are creating an environment favoring defense. Even if China attempts an amphibious or blockade operation against Taiwan, the U.S.'s advanced defense system, utilizing precision long-range strike capabilities, intelligence, surveillance, and reconnaissance (ISR) capabilities, and autonomous weapon systems, would make it difficult for China to achieve its military objectives. Even if partially achieved, it would come at an immense cost of war. Conversely, China, through its long-range precision strike capabilities, could inflict significant damage on key U.S. command centers and air bases within the first island chain, thereby considerably limiting the operational capabilities of U.S. forces targeting mainland China in a contingency. In other words, in the event of a conflict between the U.S. and China in the Taiwan Strait, either side would face extremely high costs for launching an attack on the other (Fravel and Heginbotham 2024). Additionally, drawing attention to the fact that generative AI requires vast amounts of data to function properly, it is argued that AI-enhanced unmanned autonomous weapon systems are more advantageous for defense, as the defending nation possesses more detailed information about its own terrain and military installations than the adversary (King 2024).

Third, the nuclear-AI nexus has the characteristic of enhancing both 'first strike capability (or counterforce capability)' and 'second strike capability (or countervalue capability)'. Improvements in ISR capabilities, cyber attacks and information disruption against enemy nuclear command and control (NC2) facilities, and increased precision of conventional counterforce strikes enhance first strike capability by destroying the adversary's nuclear capabilities. Conversely, enhanced early warning and situational awareness capabilities, cyber defense, defensive use of drone swarms, and automated retaliatory systems for nuclear weapons contribute to improving second strike capability in response to and defense against enemy nuclear attacks (Kim Yang-gyu 2024b). To correctly understand whether first or second strike capability contributes to offense or defense, it is necessary to consider that Mutual Assured Destruction (MAD) is established when both sides possess credible second strike capabilities, which enhances strategic stability between those nations. That is, second strike capability enhances strategic stability by increasing the credibility of MAD, while first strike capability intensifies instability by seeking to destroy it. In this context, the fact that AI-integrated nuclear weapon capabilities are enhancing both second strike capability and the first strike capability that neutralizes it makes it difficult to definitively state whether the introduction of AI technology leads to an advantage in offense or defense, even in terms of nuclear balance.

Based on the above, the common points highlighted in recent research on the military use of AI can be summarized into three points. First, AI technology, from any perspective, is difficult to consider as inherently providing an advantage in offense or defense. This is a natural consequence given the technology's general-purpose nature and its role as a force multiplier. Scholars who study military technological innovation emphasize that, considering the action-reaction between the military use of new technologies and the adversary's adaptation, a definitive answer to the offense-defense balance problem cannot be reached based on a single technological variable (Biddle 2023). The general-purpose nature of AI further highlights this action-reaction dynamic.

Second, there is the tension between efficiency and stability in AI technology. To counter the emergence of offensive advantages due to technological change and enhance AI-based defensive capabilities, it is necessary to limit human intervention and control (human out of the loop) and increase machine autonomy. However, while this reduces warfare costs, it increases the possibility of 'unintended escalation.' Conversely, the more responsibly and safely AI is utilized, the less effective its integration into military power becomes, and the resulting decrease in efficiency leads to outcomes contrary to the original purpose of military AI application. In particular, if the adversary abandons safe AI application methods and opts for strategies focused on efficiency that rely on machine autonomy, while the defending nation emphasizes responsible AI use and strengthens human commanders' intervention and control, it may face a situation where the efficiency of the adversary's AI-cyber, AI-autonomous weapon systems, and AI-nuclear weapon systems overwhelms the defending nation's AI capabilities.

Third, while it is difficult to definitively conclude whether the military use of AI results in an exclusive advantage for offense or defense, and paradoxes exist due to the efficiency-stability trade-off, many studies on military capabilities, such as AI-enhanced cyber attack capabilities or AI-integrated autonomous weapon systems, predict the advent of an 'era of offensive advantage' following the military application of AI. Particularly, perceptions among U.S. and Chinese political elites suggest a widespread expectation that AI will confer an offensive advantage when integrated with existing military capabilities, exhibiting a pattern similar to the 'cult of the offensive' phenomenon prior to World War I (Selden 2024).

2. Interpretation of the Global Security Reality: Perception of Inevitable Conflict and the Cult of the Offensive

Why do policymakers fail to accurately perceive the objective reality of military technological changes and instead anticipate an era of offensive advantage? The phenomenon of a distinct gap between the reality of the offense-defense balance and its perception has been a recognized issue since Jervis's early research, which emphasized the importance of this variable. Jervis presents four theoretically possible worlds centered on the 'offense-defense balance' and 'offense-defense differentiation,' explaining that a 'doubly stable' security environment, free from the security dilemma, is formed when there is a 'defensive advantage' and 'offense and defense technologies are distinguishable.' Jervis identifies the first decade of the 20th century, prior to World War I, as perhaps the only period in human history when such stable conditions existed. He argues that if European powers had properly understood the implications of machine guns and trench warfare, they would have chosen different security policies, and World War I would not have occurred. Nevertheless, almost all strategists in the major powers at the time believed that offense had the advantage and that offensive and defensive technologies were indistinguishable. This 'perception' overwhelmed the physical 'reality' (Jervis 1978, 211-214).

Recent research explains that policymakers fail to properly interpret the realities of change in military technology because "a mix of ambiguous technological advances and growing sense of inevitable conflict" forms a vicious cycle of perception that reinforces itself (Selden 2024, 6). The study reports these findings after sampling and analyzing discussions on the military use of AI from 2014 to 2022, including documents such as interviews and speeches by the U.S. President and Secretary of Defense, official Department of Defense strategy documents, reports from various government committees, and congressional statements, as well as documents like China's security strategies, government reports, interviews with President Xi Jinping, speeches by high-ranking Chinese Communist Party officials, and commentaries from Chinese state-run media such as Xinhua News Agency. Furthermore, it explains that policymakers in the United States and China exhibited perceptions of the possibility of conflict between the two countries as shown in [Figure 1]. The section where the blue solid line descends below the dotted line represents phases where the perception of inevitable conflict between the two countries is overwhelming.

[Figure 1] Conflict/De-escalation Discussions Among U.S. and Chinese Policymakers

In the case of the United States, perceptions of China were strongly adversarial in the early period of Trump's first term, then turned positive from 2018, but a perception of inevitable conflict with China surged with the inauguration of the Biden administration in 2021. China's perception of the U.S., which worsened from 2019, has not recovered since then, and the perception that conflict and confrontation between the two countries are unavoidable has become mainstream. If these perceptions become entrenched among U.S. and Chinese political elites, and the ambiguous effects of technology on the offense-defense balance persist, it is highly probable that the U.S. and Chinese leaderships will perceive that the military utilization of AI will usher in a world of offensive advantage, assuming the observations of this study are accurate. This argument is highly persuasive, aligning with existing research on motivated reasoning, which suggests that policymakers, when 'feeling compelled to act,' exhibit a tendency to intentionally ignore all signals regarding the adversary's will and intentions by employing denial, selective interpretation, and other psychological mechanisms (Lebow 1981). The belief that one must fight the adversary leads to a desire to believe that technological advancements have created conditions favorable to offense.

III. U.S.-China Arms Race in 2025 and South Korea's Defense Policy

When leaders believe they are in a military-technological environment favoring offense, they tend to expect greater gains from war, leading to intensified arms races and solidified alliance orders. The security dilemma worsens, and the likelihood of conflict increases even among status quo powers (Jervis 1978, 189-190). Just as the leaders of European powers did before World War I, the power of perception can override physical reality. Therefore, we arrive at the conclusion that the perceptions of leaders interpreting the reality of AI's military application may be more important than the substantive truth of the changing reality itself. In this context, Trump's return, South Korea's domestic political turmoil, and the crisis of political leadership must be taken as serious issues.

Of course, the probability of a direct military conflict between the U.S. and China within 2025 is quite low. The conclusion of the Biden-Xi summit held during the Asia-Pacific Economic Cooperation (APEC) summit in Lima, Peru, in November 2024 was that communication channels between the two countries must be continuously maintained to 'manage competition responsibly and prevent it from veering into conflict or confrontation' (The White House 2024). Furthermore, among the five principles of diplomacy emphasized by Foreign Minister Wang Yi at the '2024 International Situation and China's Diplomacy' forum on December 17, the first principle was 'peace' (Ministry of Foreign Affairs, The People’s Republic of China 2024). However, considering the changes in military technology discussed earlier and the perceptions of U.S. and Chinese leaderships towards them, this situation is more akin to a 'calm before the storm' rather than a harbinger of 'peace for our time.' At the Lima U.S.-China summit, the U.S. clearly expressed its displeasure regarding Taiwan, freedom of navigation, China's support for Russia's defense industry, and Chinese cyber attacks on U.S. civilian infrastructure. The second principle, 'unity,' and the fourth principle, 'justice,' emphasized in Foreign Minister Wang Yi's speech, highlighted the strengthening of solidarity among the Global South, BRICS, and China, and an 'equitable' international order. Major U.S. think tanks are publishing reports outlining the possibility of U.S.-China military conflict in the Taiwan Strait and the South China Sea in the short to medium term and proposing U.S. responses (Cancian, Cancian, and Heginbotham 2023; O’Hanlon 2024; Sisson and Patt 2024).

Currently, as both the U.S. and China have agreed to maintain human control over decisions regarding the use of nuclear weapons, the risk of unintended nuclear war stemming from the nuclear-AI nexus competition is relatively low. However, with China's plan to increase its nuclear warhead production to 700-1,500 by 2035 (Kristensen, Korda, Johns, and Knight 2024) and the U.S.'s policy direction to enhance its tactical nuclear capabilities, including the production of low-yield tactical nuclear warheads W76-2 for submarine-launched ballistic missiles (SLBMs) and the improved low-yield tactical nuclear bomb B61-12 for F-35s, classic vertical nuclear proliferation and arms race between the U.S. and China will continue. Furthermore, in the absence of specific agreements on the development and military use of AI technology, beyond the general principle of 'carefully considering potential risks and developing responsibly,' competition between the U.S. and China is expected to intensify over AI-autonomous weapon systems and AI-cyber attack/defense capabilities.

Last week, Jensen Huang, CEO of Nvidia, predicted that the commercialization of quantum computers would take another 15-20 years, causing a stock market plunge in quantum computing-related stocks. More significant than the stock price drop is the fact that if the era of quantum computing arrives more slowly than expected, the short- to medium-term importance of AI technology will further increase. When the period of martial law and impeachment concludes and South Korean leadership enters a phase of stability, it is necessary to establish defense policy directions that consider long-term challenges. Even if the situation is 'the eve of a storm' rather than 'peace for our time,' the window of opportunity for South Korea still remains open during this 'decisive decade.'

First, it is necessary to strategically leverage the U.S.-China arms race. The belief that one is in an offensive-advantage environment provides a strong incentive to strengthen cooperation among allies and increase investment in offensive assets. This is likely to manifest as an independent trend within the U.S. Department of Defense, separate from Trumpism. Therefore, South Korea should actively utilize this to further strengthen security cooperation with the U.S., a leader in AI technology, based on new technologies. Centered around the Strategic Command established last October, South Korea needs to develop doctrines for the integrated utilization of its strategic assets, enhance Kill Chain capabilities, establish operational plans and conduct joint exercises for U.S.-ROK integrated extended deterrence and Conventional-Nuclear Integration (CNI), strengthen U.S.-ROK-Japan security cooperation, and, above all, make efforts to enhance cooperation in a wide range of areas including the U.S. Department of Defense's development of operational concepts for the military use of AI, military structure reforms, weapon system development and procurement, and talent cultivation. Of course, unlike the Biden administration, the ROK-U.S. alliance under a second Trump administration is likely to become more transactional. In that case, it will be crucial to carefully monitor changes in the pricing of technology cooperation and transfer and respond proactively.

Fundamentally, it is more important to develop the foundational capabilities for South Korea's AI technology that can be utilized militarily, rather than focusing solely on developing offensive capabilities. To build an AI-powered, technologically advanced military and a hybrid manned-unmanned combat system, as emphasized by 'Defense Reform 4.0,' high-quality and large-scale military data are needed for operational planning, in addition to indigenous AI technology infrastructure and semiconductor capabilities. If South Korea sets its national AI strategy by specializing in small or ultra-small AI models, considering the scale of its data and server market, rather than adopting the Large Language Model (LLM) approach pursued by the U.S. (Bae Young-ja 2025), then dedicated research and active investment into what form a Korean-style AI military model, utilizing this, should take are absolutely essential. ■

References

Kim Yang-gyu. 2024a. “Global Military Order in 2024 and Korea: South Korea's Security Policy in the Era of Offensive Advantage Driven by the Revolution in Accuracy and Transparency.” EAI New Year Special Commentary Series. January 5.https://www.eai.or.kr/new/ko/pub/view.asp?intSeq=22294(Accessed: January 14, 2025.)

______. 2024b. “The AI-Nuclear Nexus and the Outlook for the Global Military Order.” EAI AI and New Civilization Standards Special Report. September 6.https://www.eai.or.kr/new/ko/pub/view.asp?intSeq=22705&board=kor_special(Accessed: January 14, 2025.)

Park Sung-min. 2024. “Victor Cha: 'Prolonged Korean Chaos, Worst-Case Scenario for Trump's Second Term ROK-U.S. Alliance'.” Yonhap News Agency. December 13.https://www.yna.co.kr/view/AKR20241213002700071(Accessed: January 14, 2025.)

Bae Young-ja. 2025. “AI Technology Competition and Global Politics in 2025: South Korea's Response Strategy.” EAI New Year Special Commentary Series. January 13.https://eai.or.kr/new/ko/pub/view.asp?intSeq=22684(Accessed: January 14, 2025.)

Son Yeol, Kim Yang-gyu, Lee Dong-ryul, Lee Seung-ju, Jeon Jae-seong, and Ha Young-sun. 2023. “Between 'Managed Competition' and 'Securing Development Space': The 2023 APEC U.S.-China Summit Seeking Cooperation.” EAI Special Report. November 20.https://www.eai.or.kr/new/ko/pub/view.asp?intSeq=22234(Accessed: January 14, 2025.)

Lee Jae-rim. 2025. “[Trump's Second Term Inauguration] Everyone Rushing for Recognition... Attempts to Improve Relations with Trump Continue.” Yonhap News Agency. January 12.https://www.yna.co.kr/view/AKR20250111011400087(Accessed: January 14, 2025.)

Jeon Jae-seong. 2024. “AI-Based Autonomous Weapon Systems, Cognitive Warfare, and Changes in the Military Security Order.” EAI AI and New Civilization Standards Special Report. September 6.https://www.eai.or.kr/new/ko/pub/view.asp?intSeq=22706&board=kor_special(Accessed: January 14, 2025.)

______. 2025. “Trumpism's Diplomatic Strategy and the Future of World Order: ROK-U.S. Relations.” EAI New Year Special Commentary Series. January 3.https://www.eai.or.kr/new/ko/pub/view.asp?intSeq=22678(Accessed: January 14, 2025.)

Biddle, Stephen. 2001. “Rebuilding the Foundations of Offense-Defense Theory.”The Journal of Politics 63, 3: 741–774.

Cancian, Mark F., Matthew Cancian, and Eric Heginbotham. 2023. The First Battle of the Next War: Wargaming a Chinese Invasion of Taiwan. Washington, DC: Center for Strategic and International Studies.

Fravel, M. Taylor, and Eric Heginbotham. 2024. “Envisioning a Stable Military Balance in 2034.” In U.S.-China Relations for the 2030s: Toward a Realistic Scenario for Coexistence, eds. Christopher S. Chivvis, 79–90. Washington, DC: Carnegie Endowment for International Peace.

Horowitz, Michael C. 2018. “Artificial Intelligence, International Competition, and the Balance of Power.” Texas National Security Review 1, 3: 36-57.

Jacobsen, Jeppe T., and Tobias Liebetrau. 2023. “Artificial Intelligence and Military Superiority: How the ‘Cyber-AI Offensive-Defensive Arms Race’ Affects the US Vision of the Fully Integrated Battlefield.” In Artificial Intelligence and International Conflict in Cyberspace, eds. Fabio Cristiano, Dennis Broeders, François Delerue, Frédérick Douzet, and Aude Géry, 135–156. London: Routledge.

Jervis, Robert. 1978. “Cooperation Under the Security Dilemma.” World Politics 30, 2: 167-214.

King, Anthony. 2024. “Robot Wars: Autonomous Drone Swarms and the Battlefield of the Future.” Journal of Strategic Studies 47, 2: 185–213.

Kristensen, Hans M., Matt Korda, Eliana Johns, and Mackenzie Knight. 2024. “Chinese Nuclear Weapons, 2024.” Bulletin of the Atomic Scientists 80, 1: 49–72.

Lebow, Richard Ned. 1981. Between Peace and War: The Nature of International Crisis. Baltimore: Johns Hopkins University Press.

Lieber, Keir A., and Daryl G. Press. 2017. “The New Era of Counterforce: Technological Change and the Future of Nuclear Deterrence.” International Security 41, 4: 9–49.

Ministry of Foreign Affairs, The People’s Republic of China. 2024. “Riding the Trend of the Times with a Strong Sense of Responsibility.” December 17. https://www.fmprc.gov.cn/eng/wjbzhd/202412/t20241218_11497818.html (Accessed January 14, 2025)

O’Hanlon, Michael E. 2024. “Could the United States and China Really Go to War? Who Would Win?” Brookings, August 15, 2024. https://www.brookings.edu/articles/could-the-united-states-and-china-really-go-to-war-who-would-win/ (Accessed January 14, 2025)

Schneider, Jacquelyn, and Julia Macdonald. 2024. “Looking Back to Look Forward: Autonomous Systеms, Military Revolutions, and the Importance of Cost.” Journal of Strategic Studies 47, 2: 162–184.

Selden, Zachary. 2024. “A New ‘Cult of the Offensive?’ Elite Perceptions of Artificial Intelligence in Military Affairs in the United States and the People’s Republic of China.” Foreign Policy Analysis 20, 4.

Sisson, Melanie W., and Dan Patt. 2024. “After Primacy: US Military Options in Contemporary East Asia.” Brookings, December 20, 2024. https://www.brookings.edu/articles/after-primacy-us-military-options-in-contemporary-east-asia/ (Accessed January 14, 2025)

Waltz, Kenneth Neal. 1979. Theory of International Politics. New York: McGraw-Hill.

The White House. 2024. “Readout of President Joe Biden’s Meeting with President Xi Jinping of the People’s Republic of China.” November 16. https://www.whitehouse.gov/briefing-room/statements-releases/2024/11/16/readout-of-president-joe-bidens-meeting-with-president-xi-jinping-of-the-peoples-republic-of-china-3/ (Accessed January 14, 2025)

■ Kim Yang-gyu_Senior Research Fellow, East Asia Institute; Lecturer, Department of Political Science and International Relations, Seoul National University.

■ Managed and Edited by: Park Han-soo_Research Fellow, EAI

    Inquiries: 02 2277 1683 (ext. 204) hspark@eai.or.kr