En
- Research Outputs Think Tank Reports Academic Papers Current Affairs Commentary Research References Academic Works
- Academic Exchange International Conferences Domestic Conferences Domestic and International Exchange
- Speaking Out Externally Podcast Newsletter Media Interviews Communication Reference
- Center Team Academic Committee Research Team Special Experts Visiting Scholars Administrative Team
- Talent Development International Youth Exchange Strategic Youth Tsinghua University Student Association for Global Strategy Studies Tsinghua University Student Association for China-US Cultural Exchange Overseas Practical Activities Doctoral Researcher
Research Outputs
Current Affairs Commentary
Sun Chenghao: How Southeast Asia Can Seize the Initiative Amid China-US AI Competition
2026-11-04
Sun Chenghao Associate Research Fellow, Center for Strategic and Security Studies, Tsinghua University
Southeast Asia's choices in the AI field will be shaped by China-US rivalry for a long time to come, but countries can still strengthen their ability to make choices through their own investments and regional cooperation. If China-ASEAN cooperation can support these efforts, it can provide more practical help for regional countries to calmly respond to great-power competition and further solidify the foundation for cooperation between the two sides.
Recently, I traveled to Hanoi, Vietnam, to attend a seminar on artificial intelligence, strategic competition and regional security, where I exchanged views with many representatives from Southeast Asia. The topic of China-US strategic competition ran through almost the entire day's discussions, but what impressed me most was how participants thought about their own countries' situations. They hope to use advanced technology to accelerate their own development, but they also worry about how much autonomy their countries can retain once public services, critical infrastructure and security decisions increasingly depend on foreign technology. Southeast Asia's choices in the AI field simultaneously affect the balance of development, security and foreign relations.
In the past, Southeast Asian countries often avoided excessive dependence on any single major power by cooperating with several of them; the development of AI makes this balancing strategy harder to implement. Because many systems require continuous access to external computing power, model updates and technical support, dependence deepens gradually with use. Once technology enters government administration and critical infrastructure, supply disruptions may affect the normal operation of the state. At that point, even if a government wishes to adjust its foreign policy, it must consider whether existing technology can continue to be used, and technological dependence may thus constrain diplomatic choices.
US competition with China adds political risk to this dependence. In July 2025, Malaysia announced licensing controls on the export, transshipment and transit of US-origin high-performance AI chips. Great-power technology controls have thus entered the development agendas of regional countries. When attracting investment and building digital infrastructure, governments need to consider both whether technology can be obtained reliably and what diplomatic pressure cooperation will bring. The greater the uncertainty in China-US relations, the harder it is for Southeast Asia to consider economic development separately from security concerns.
▲On August 17, the Thailand National Astronomical Research Institute exhibit at the Thailand National Science and Technology Fair in Nonthaburi, Thailand. (Photo: Xinhua)
However, Southeast Asian countries will not all make the same choices because of this. Singapore, with its own advantages, is better placed to participate in technology R&D and international rule governance; Malaysia and Vietnam are more focused on digital infrastructure construction and upgrading local industries; countries with weaker digital foundations are still building up basic talent reserves and facilities. Countries also differ in their political and security relations with China and the US, so their assessments of external risks naturally differ. Therefore, each country will cooperate with different partners in different fields according to its stage of development and security needs.
This variation also means that one cannot judge a country's diplomatic orientation merely by which country's technology it uses. A country can use Chinese models to improve public services, conduct security cooperation with the US, and learn governance experience from other partners. These choices are not necessarily in conflict, but the government needs to coordinate the conditions attached to different cooperation arrangements. Partners such as Europe, Japan, South Korea and India can help regional countries expand their options, but they can hardly eliminate dependence on key Chinese and US technologies. Southeast Asian countries need to identify which cooperation arrangements can complement one another and which arrangements will hinder future cooperation with other countries.
To reduce these constraints, Southeast Asian countries also need to turn external technology into capabilities they can master themselves. Governments need to integrate AI development with education, research and industrial policy, continuously cultivate talent, and improve the basic conditions required for technology use. If AI is applied only in a few sectors and most of society still lacks the ability to use and judge technology, the state will find it difficult to fully capture development gains and will be more prone to long-term dependence on external support; such capacity building is hard to accomplish by merely introducing a few large projects.
Most Southeast Asian countries can hardly invest funds comparable to the scale of China and the US to develop globally leading models, but they can develop their own expertise around their national languages and public service needs. The SEA-LION project promoted by AI Singapore, for example, develops models around Southeast Asian languages and cultures; different versions have used US companies' Llama and Gemma as well as China's Tongyi Qianwen (Qwen). Such attempts show that using foreign technology and building domestic capacity can advance at the same time. Only when countries have the ability to evaluate and improve technology can they decide which technologies suit them and which uses need to be restricted.
This kind of judgment capability is especially important in the security field. When AI is used for maritime monitoring, intelligence analysis and cyber defense, bringing in external technology may also mean accepting the external system's framework for judging threats. If governments cannot independently examine these judgments, they may unknowingly be influenced by technology providers. In regional disputes and crises, how to understand the other side's intentions and when to take action should still be decided by each country according to its own interests. To this end, countries need to retain the ability to verify information, correct errors and suspend use of systems, so that technology serves national decision-making.
But countries' own investments alone may not solve these problems. Among Southeast Asian countries, there are significant gaps in talent reserves and infrastructure levels, and weaker countries find it harder to evaluate external technology and harder to participate in international rule discussions. If ASEAN can organize joint research, share testing resources and conduct personnel training, it can reduce the burden of duplicate construction among countries. Regional cooperation can thereby help more members master technology, and also enable ASEAN to put forward clearer common demands when negotiating with major powers, changing the previous passive situation in which countries could only accept external conditions individually.
The regional AI safety network that ASEAN is advancing can provide a platform for these efforts. In September this year, ASEAN digital senior officials and heads of telecommunications regulators continued to discuss the establishment of this network and emphasized cooperation with external partners. ASEAN countries have different laws and levels of development, so it is difficult to formulate completely uniform rules in the short term, but they can first make common arrangements for cross-border risks, for example by agreeing on which incidents need to be reported and how to coordinate after problems arise. This can both preserve each country's right to formulate domestic policies and provide clear rules for different partners to participate in regional cooperation.
For China, understanding Southeast Asia's desire to independently decide its development direction helps make cooperation more solid. China and ASEAN can link technological cooperation with each country's development plans, increase investment in talent training, public services and governance research, and also consider providing more support to weaker members. When discussing regional standards and safety rules, ASEAN countries' opinions should be fully incorporated so that cooperation better meets local needs. Enabling regional countries to independently control technology applications and participate in rule-making helps dispel doubts and build a more lasting foundation of mutual trust.
▲In September 2025, the 22nd China-ASEAN Expo and China-ASEAN Business and Investment Summit opened in Nanning, Guangxi. (Source: Xinhua News Agency)
Such cooperation also requires a relatively stable environment in China-US relations. My experience participating in the China-US AI Track II dialogue made me feel that even amid fierce competition, the two countries still need to maintain communication around major security incidents. If a technological accident or a cross-border attack launched with AI is misjudged as a state-level hostile action, it could very likely rapidly escalate tensions in bilateral relations. Southeast Asia will also be affected by these risks and should participate in relevant discussions and raise its own security concerns. If China-US cooperation on risk issues can be opened to more regional partners, it can also reduce the impact of competition on normal cooperation.
From observations of the Hanoi meeting, Southeast Asia's choices in the AI field will be influenced by China-US competition over the long term, but each country can still strengthen its ability to make choices through its own investment and regional cooperation. ASEAN has already issued the "2025–2030 Responsible AI Roadmap," which sets an overall framework for regional AI development and provides a layered implementation path for member states to advance AI development in line with their own conditions, and countries in the region have already begun to increase investment. They need to decide development priorities based on their national conditions, strive to obtain key technologies through different channels, and retain the ability to make independent judgments on security issues. If China-ASEAN cooperation can support these efforts, it can provide more practical help for regional countries to calmly respond to great-power competition and further solidify the foundation for cooperation between the two sides.
This article was first published on October 26, 2026, in "China-US Focus."
More Recommendations
Yu Xiang: Global Financial Stability Requires Joint Safeguarding by China and the US
Wen Jing: Gulf Countries' New Perceptions of China-US Relations and Policy Adjustments
Jiang Feng | Climate Cooperation: Opening New Horizons for China-Europe Relations
Sun Chenghao: The NATO Dilemma Behind the US "Strategic Retreat" Anxiety
Substack
Copyright ©2019 Center for Strategic and International Studies, Tsinghua University All rights reserved
Email: ciss@mail.tsinghua.edu.cn
Postal code: 100084
Address: Room 428A, Mingli Building, Tsinghua University, Haidian District, Beijing