[International Politics in the Age of AI] IV. International Cooperation on Artificial Intelligence (AI): Focusing on the Quad, AUKUS, and Middle Power Alliances

I. Introduction

The second Trump administration, inaugurated on January 20, 2025, unlike its first term, recognizes Artificial Intelligence (AI) as central to maintaining global hegemony. The first Trump administration only managed to release limited AI policies towards the end of its term. In contrast, the second Trump administration placed the AI agenda at the forefront during the election campaign and, just three days after its inauguration, abolished the Biden administration's Executive Order 14110 on AI and promulgated a new executive order with a pro-business orientation. This action clearly signaled a departure from the existing regulatory and risk management-centric approach, prioritizing private sector autonomy and innovation promotion. Subsequently, on July 23, 2025, the "Winning the AI Race: American AI Action Plan" (hereinafter referred to as AI Action Plan) was officially announced, comprehensively presenting the AI policy for the second term.

The AI Action Plan consists of three pillars, with the third pillar (Lead in International AI Diplomacy and Security – Export American Allies and Partners) directly related to AI diplomacy and cooperation. The core objective is to establish a US-centric security and technology ecosystem as a global standard by expanding the adoption of American AI products and solutions among allies and partners. However, while this directive professes cooperation, it strongly compels or drives the adoption of American hardware and software. In 2017, when the US sanctioned China's Huawei, it demanded the exclusion of Chinese products; this time, however, it goes beyond sanctioning Chinese products to actively press for the adoption of American ones. In other words, the intention is to solidify American AI hegemony through the Americanization of international standards and leadership in technology diffusion. In this context, the US is seeking to strengthen AI alliances centered around itself, and in response, China and Russia are also broadening the scope of their AI cooperation.

II. US AI Cooperation: Quad and AUKUS

The US has expressed its intent to deepen AI alliance/cooperation diplomacy in the AI Action Plan. AI agendas are already being addressed across various multilateral platforms led by the US, including the Global Partnership on AI (GPAI), the Quad, the second pillar of AUKUS, the Five Eyes, NATO, the Indo-Pacific Economic Framework (IPEF), and the Summit for Democracy.[1]The core among these are the Quad and AUKUS.

1) Current Status

The US is utilizing the Quad (US, Japan, Australia, India) and AUKUS (US, UK, Australia) as hubs for AI and advanced technology cooperation. The Quad is already operating applied cooperation projects such as AI-ENGAGE (agriculture) and BioExplore (biodiversity), and is advancing agenda-specific linkages through initiatives like the Working Group on Critical and Emerging Technology launched in 2021. The 'Quad Tech Network' at the second track level has also been operational since December 2020, expanding the foundation for continuous information exchange and joint research. Participants include the National Security College at the Australian National University, India's Observer Research Foundation, Japan's National Graduate Institute for Policy Studies, and the US's Center for a New American Security. The Quad countries operate a 'Senior Cyber Group' that facilitates regular cooperation among experts from the four nations. This group pursues joint responses and technological cooperation with the goals of ensuring the security of critical infrastructure, strengthening cybersecurity capabilities, and enhancing resilience against cyber threats.[2]

The Quad has the advantage of presenting a normative alternative to China's technological authoritarianism model, given that all four countries share the common denominator of democracy. Particularly from the perspective of cyber-AI linkage, it aims to prevent the misuse of AI, such as surveillance, censorship, and disinformation, and to secure a technological ecosystem based on openness, accessibility, and safety as a common goal. The Quad Foreign Ministers' Meeting held in July 2025 adopted cooperation on Information and Communication Technology (ICT) infrastructure, including AI, semiconductors, technology standards, and cybersecurity, as a key agenda item and designated undersea cable security as a priority cooperation area. The US is preparing measures to restrict US connections for undersea cables that include Chinese technology.[3]Furthermore, the four countries possess complementary expertise in AI. The US has high expertise in machine learning and natural language processing, Australia in linguistics and theoretical computer science, India in data mining and data science, and Japan in human-computer interaction (HCI). These distinct comparative advantages provide a foundation for creating synergistic effects through cooperation.

Meanwhile, AUKUS is a cooperation among the US, UK, and Australia, key members of the Five Eyes. Traditionally, the Five Eyes has led the joint development of dual-use technologies and the sharing of sensitive information among the US, UK, Australia, Canada, and New Zealand since the Cold War. Based on shared political, cultural, and linguistic identities, these five nations have expanded their cooperation to include cybersecurity and advanced technology protection mechanisms today. However, concerns about the weakening cohesion of the Five Eyes have been raised following the sharp deterioration of US-Canada relations due to the imposition of high tariffs on Canada immediately after the inauguration of the second Trump administration and President Trump's remarks about Canada being 'the 51st state of America.' New Zealand has been a relatively weaker link among the five countries since the 'ANZUS crisis' in the mid-1980s due to its anti-nuclear policy. In this context, cooperation among the three AUKUS countries (US, UK, Australia) within the Five Eyes framework may function as its core axis.

AUKUS proceeds on two tracks. 'Pillar 1' is cooperation for the US and UK to provide Australia with nuclear-powered submarines, and 'Pillar 2' aims for joint development and enhanced interoperability of advanced technologies among the three countries. There was considerable debate among experts regarding whether AUKUS, agreed upon during the Biden administration, would continue as planned under the second Trump administration. In early June 2025, there were mentions of a review of AUKUS by the US side. However, in mid-June, during the summit meeting between the US and UK on the sidelines of the Group of Seven (G7) summit, President Trump and Prime Minister Starmer confirmed the continued pursuit of AUKUS. Subsequently, on October 20, President Trump met with Australian Prime Minister Anthony Albanese at the White House and announced his intention to expedite the procedures for providing nuclear-powered submarines to Australia. Therefore, the progress of 'Pillar 1' is expected to enter a stable phase.

In contrast, 'Pillar 2' consists of eight technology and capability areas, with cooperation in AI, cyber, quantum computing, and undersea technologies prioritized for operation by 2025, followed by an expansion of cooperation to hypersonics, electronic warfare, innovation, and information sharing. The scope of cooperation, as well as the participating countries, is expected to expand. The AUKUS trilateral joint statement on September 17, 2024, explicitly mentioned consultations on 'Pillar 2' with Canada, New Zealand, and South Korea. In September 2025, Japan participated for the first time in AUKUS 'Pillar 2' activities to discuss cooperation in unmanned aerial vehicle (UAV) AI research, which aligns with the recent strategic environment where the US and Europe are deepening defense industry cooperation with Japan to counter China's military expansion. The UAV AI research between the AUKUS countries and Japan is considered a technological foundation that could potentially be applied to the next-generation fighter jet joint development project that Japan is pursuing separately with the UK and Italy.[4]

2) Constraints and Challenges

While the Quad's strengths lie in its symbolism as a democracy alliance and the openness of its Plus (+) format, India's strategic reservations are a constant factor. As AI becomes a central pillar of US-China technological hegemony, it remains uncertain whether India will actively utilize the Quad as a platform for AI cooperation. Recent discord between the US and India, stemming from US imposition of high tariffs and India's purchase of Russian crude oil, further fuels these concerns. Fundamentally, unlike the Five Eyes countries, India has a limited degree of shared identity with the US due to cultural and historical differences.

The AUKUS 'Pillar 2' faces constraints due to the US export control regime, particularly the International Traffic in Arms Regulations (ITAR), which can delay or block the transfer of technology and equipment. The UK incurs annual costs equivalent to approximately 1% of its defense budget due to ITAR, and Australia also experiences thousands of permit delays each year.[5] In response, some companies, such as Boeing and Anduril, are attempting to circumvent ITAR restrictions through 'Research and Development (R&D)' conducted locally in Australia.[6] Although the US Department of State implemented some relaxations for the UK and Australia in August 2024, strategically sensitive areas such as nuclear propulsion and quantum navigation remain under regulatory control. For AI cooperation through AUKUS to make substantial progress, further relaxation of institutional barriers, increased efficiency in trilateral defense industry collaboration, and bold relaxation of controls on basic research are necessary. However, as mentioned earlier, the third pillar of the US AI Action Plan includes intentions to strengthen export controls on computing resources alongside expanding US AI exports, indicating that institutional constraints still remain in making AUKUS 'Pillar 2' a core platform for US AI cooperation.

III. AI Cooperation among China, Russia, and India

In response to the US, China is also discussing AI agendas within networks such as the Belt and Road Initiative (BRI), the Shanghai Cooperation Organisation (SCO), and BRICS. For instance, President Xi Jinping emphasized at the SCO Summit held in Tianjin, China, on September 1, 2025, that member states should expand cooperation in the field of AI.[7] He also stated that AI development should be achieved through mutual linkage and joint efforts, urging a move away from 'Cold War mentality' towards a cooperative approach. Furthermore, China is strengthening its efforts in discussions within multilateral platforms such as the UN. Following the release of the 'Global AI Governance Initiative' in 2023, at the World Artificial Intelligence Conference (WAIC) held on July 26, 2025, it proposed the establishment of a World AI Cooperation Organization and announced the 'Global AI Governance Action Plan,' attempting to establish alternative norms.

China seeks to expand its international influence through AI cooperation, deploying a multi-layered global cooperation strategy.[8] First, it is promoting digital transformation in developing countries by strengthening the 'Digital Silk Road' program, which supports the construction of digital infrastructure and AI technologies. Second, by making its developed AI models and technologies freely available, it is creating an open-source foundation for global developers to participate, thereby aiming to lead standards and influence within the international AI ecosystem. Third, it is actively expanding technological cooperation with emerging economies in regions such as Africa and Southeast Asia, strengthening its presence in new markets based on strategic partnerships.

Russia is actively pursuing deepened cooperation with BRICS and other emerging economies, including China, to expand its influence in the AI field. Russian President Vladimir Putin has expressed his intention to jointly develop AI with BRICS and other emerging nations, emphasizing the need for an international AI cooperation platform. He further suggested that the formation of an international union involving AI-related academic societies, institutions, and R&D organizations from BRICS countries could significantly accelerate joint research and technological cooperation in AI, presenting a concrete vision for cooperation.[9]

Meanwhile, the Global South, including India, is focusing on building a third force that is not unilaterally subordinate to the norm competition centered on the US and China, by expanding participation in GPAI. However, even if many countries in the region, such as India, hasten to build an AI ecosystem under the banner of securing 'Sovereign AI,' the gap with the US and China in core elements such as capital, technology, data, and talent remains large. For this reason, the necessity of multilateral solidarity among middle powers, transcending individual efforts, is emerging.

IV. Possibility of Middle Power AI Solidarity?

US allies and security partners face pressure to choose between 'Sovereign AI' and the adoption of standards established (and being established) by the US. Furthermore, even if they choose to adopt US products due to concerns about the security vulnerabilities of Chinese products, issues regarding the security and reliability of US products may also arise. In other words, US allies and security partners are facing the dilemma of choice and priority investment at the intersection of sovereignty, standards, and security.

However, due to the Trump administration's second term's "America First" policy and disregard for alliances, the cohesion of US-led AI research and investment networks is low. The US pressured NATO members to increase defense spending and hinted at the possibility of withdrawing from NATO, and Vice President Vance's coercive attitude at the Munich Security Conference in February 2025 triggered backlash from allies. Additionally, reports that Secretary of Defense Pete Hegses ordered a temporary halt to cyber operations against Russia to avoid disrupting the atmosphere for negotiations between Russia and Ukraine heightened the concerns of allies. This overlapped with former National Security Advisor Michael Waltz's designation of only China and Iran as targets for cyber deterrence, excluding Russia, amplifying the distrust among allies.[10] As mentioned earlier, immediately after the inauguration of the Trump administration's second term, the US imposed high tariffs on Canada and President Trump's taunts towards Canada with the [11] In this context, attention is focused on whether middle powers in the region can pursue middle power AI solidarity while simultaneously implementing individual self-strengthening strategies.[12]

Indeed, an expansion of bilateral and trilateral cooperation among middle powers in the region is being observed. At the summit between Japan and India held in August 2025, Japan pledged an investment of $68 billion in India. During the summit, the two countries decided to launch the India-Japan AI Initiative, which aims to jointly develop international standards for 'trustworthy AI.' It is also possible that both countries will position themselves as leaders in global norms at the 'AI Summit' to be hosted by India in February 2026.[13] Furthermore, the Japan-Australia Foreign and Defense Ministers' 2+2 meeting held in September 2025 agreed that both countries would cooperate to build a secure and trustworthy AI framework in the Indo-Pacific region, based on the Hiroshima AI Process. South Korea and India also agreed to strengthen strategic partnerships in advanced fields such as semiconductors, AI, and clean energy through high-level talks in August 2025.

Nevertheless, the level of joint AI research and investment among Quad member countries Australia, India, and Japan is very low. Joint research among the three countries is estimated to be less than 4% of the total, and investment linkages are less than 1%, primarily due to differences in data governance, capability gaps, and misalignment of geopolitical priorities.[14] For example, India refused to sign the 'Osaka Track' proposed by Japan at the G20 Osaka Summit in 2019.

Key middle powers in Asia, including South Korea, India, Australia, Indonesia, and Singapore, have already shown high achievements in fintech and mobile technology, and are rapidly expanding AI adoption based on these strengths.[15] However, reliance on the US and China for AI infrastructure limits technological autonomy, and this gap is exacerbating inequalities in growth and quality of life. Countries with weak infrastructure are also likely to see their influence further weakened by the outflow of talent and capital. Therefore, regional solidarity is needed to jointly secure computing power, data, and talent, and to collaborate on research and product development.[16] This shared concern is also evident in President Lee Jae-myung's emphasis on the importance of digital transformation cooperation between South Korea and ASEAN at the ASEAN-led multilateral summit held in Malaysia in late October 2025.

V. Outlook and Implications for the Indo-Pacific Order

In the Indo-Pacific region, amid intensifying geopolitical and geo-economic competition between the US and China, various (mini)multilateral cooperation efforts are layered, cooperating, linking, or competing. Within the US-led security network, minilateral forums such as the Quad and AUKUS have emerged alongside bilateral alliances. However, with the deepening of the "America First" ideology after the start of Trump's second term, the US-led liberal order is being shaken, and the discourse on self-strengthening and solidarity among middle powers is expanding.

For countries in the region to exert a certain degree of normative, institutional, and functional influence beyond being passive recipients of the order, they must possess capabilities beyond those of middle powers, as well as regional leadership and the role of a broker connecting inter-regional networks. Potential middle powers in the region include Japan and South Korea (Northeast Asia), Indonesia (Southeast Asia), Australia and New Zealand (South Pacific), and India (Indian Ocean). While no single middle power can single-handedly exert decisive influence on the establishment and maintenance of regional security order, forming a 'minilateral coalition' can secure a certain leverage against the US and China. However, as this solidarity is characterized as solidarity among middle powers within the US-led security network, it remains to be seen whether it can serve as a balancing mechanism that, on one hand, reinforces the US network and, on the other hand, alleviates competition between the US and China.

In the same context, the necessity of middle power solidarity in the field of AI is clear, but due to the technical and infrastructural characteristics of AI, there are structural limitations to maintaining 'strategic ambiguity.' Amidst the overwhelming infrastructural superiority of the US and China, middle powers in the region are more likely to be required to maintain 'strategic clarity.' In other words, the possibility that middle power solidarity may become mired in the logic of bloc alignment cannot be ruled out.

As AI is expected to become a key determinant of international hierarchy in the future, if middle power AI coalitions fail to achieve tangible results, negative ripple effects will inevitably occur not only in the AI domain but also in middle power cooperation in other areas. The low level of investment and research cooperation among middle powers observed earlier supports these concerns. Under these realistic constraints faced by middle powers, the US AI Action Plan is closer to a strategic design to institutionalize a US-centric AI order rather than 'comprehensive solidarity,' and AI cooperation within the Quad and AUKUS is also likely to function as a key platform supporting US-led AI solidarity. Nevertheless, the scale of joint investment and research cooperation in the AI field among middle powers has not substantially increased, acting as a structural constraint that weakens the sustainability of middle power solidarity.■

[1]Kim, Sangbae. 2025. "US-China Artificial Intelligence Hegemonic Competition and South Korea: Transformation of International Politics and Middle Power National Strategy." 『National Strategy』 31 (2): p. 113.

[2]Mandeep Singh, “AI-driven threats heighten regional focus on cyber defense,” Indo-Pacific Defense Forum, August 12, 2025.(https://ipdefenseforum.com/2025/08/ai-driven-threats-heighten-regional-focus-on-cyber-defense/)

[3]Yonhap News. 2025. "US Pushes to Ban Use of Chinese Technology in Submarine Cables... 'Security Concerns'." July 17.(https://www.yna.co.kr/view/AKR20250717028700009)

[4]Ryu, Ho. 2025. "AUKUS Proposes Joint Research on 'AI-Equipped Drones' to Japan... Aimed at Countering China." Hankook Ilbo. March 4. (https://www.hankookilbo.com/News/Read/A2025030415540000973)

[5]Oh, Jeong-mi. 2024. "US Department of Commerce Eases Export Controls for Security Cooperation Allies AUKUS." 『Export Control Issue Report』. Strategic Trade Institute. May 3; Oh, Jeong-mi. 2025. "US Needs to Revise Export Control List for Defense Cooperation." 『Overseas Research Report 2025-22』. July 21.

[6]Shah, Rajiv. 2023. "US export rules need major reform if AUKUS is to succeed." The Strategy. Australian Strategic Policy Institute. February 16.

[7]Lee, Dong-gyu. 2025. "China's Intentions and Limitations Revealed at the 2025 Shanghai Cooperation Organization Summit." 『Issue Brief』. Asan Institute for Policy Studies. October 2.(https://asaninst.org/data/file/s1_1/222742fa858281cd85cc5db0e71d3150_UHky3TuA_56dc1199bdd7d5c4749bc1e2950bcd35a829461d.pdf)

[8]Lee, Yoon-sun. 2025. "The Political Sociology of AI: US-China AI World War to Determine Future Hegemony." The Free Daily. May 6.(https://www.jayupress.com/news/articleView.html?idxno=40197)

[9]Lee, Kyung-ah. 2024. "Putin Pursues 'AI Alliance' with China, India, etc... Challenging US Hegemony." Science Today. December 12.(https://science.ytn.co.kr/program/view.php?mcd=0082&key=202412121136489802)

[10]Shoebridge, Michael. 2025. "The ‘5 Eyes’ & Trump: problems of trust & mistrust." Strategic Analysis Australia. March 21.

[11]The Economist. 2025. "Donald Trump v the spies of Five Eyes." March 16.

[12]Horan, Hans. 2025. "Indo-Pacific Allies May Rethink US Intelligence Sharing After Gabbard, Patel Appointments." The Diplomat. March 1.

[13]Choi Yoon-jung. 2025. "‘Trump First’ and India’s Indo-Pacific Realignment." *Sejong Focus*. Sejong Institute. September 15.https://www.sejong.org/web/boad/1/egoread.php?bd=1&itm=&txt=&pg=28&seq=12397)

[14]Chahal, Husanjot, Ngor Luong, Sara Abdulla, and Margarita Konaev. 2022. "Quad AI." Issue Brief. Center for Security and Emerging Technology. May.

[15]Lim, Justin. 2025. "The AI Productivity War Has Begun — Asian Middle Powers at a Crossroads of ‘AI Infrastructure Sovereignty’!" Asean Daily News. April 16.

[16]Lim (2025).()

■ Author: Park Jae-jeok_Professor, Graduate School of International Studies, Yonsei University.

■ Responsible Editor: Lim Jae-hyun_EAI Research Fellow

    Inquiries: 02 2277 1683 (ext. 209) | jhim@eai.or.kr