Brookings: AI Safety Governance, The Southeast Asian Way

Following our moderation of the roundtable dedicated to AI policy in Southeast Asia, hosted by AI Safety Asia (AISA), we contributed to the report prepared in cooperation between AI Safety Asia and the Brookings Institution Center for Technology Innovation. The report, authored by Lyantoniette Chua (AISA), Philip Tham (AISA), and Chinasa T. Okolo (Brookings), was developed with contributions from over 30 panelists and more than 1,000 participants across six roundtable discussions held between September 2024 and March 2025. The report draws on perspectives from more than 40 contributors and reviewers representing government, academia, civil society, and the private sector across all 10 ASEAN member states and Timor-Leste.

Related

• NIST: Generative AI Profile (AI 600-1) and AI Risk Management Framework (AI RMF)

• OECD: G7 Hiroshima AI Process Reporting Framework (HAIP)

AI Governance in Southeast Asia

Southeast Asia is a region of over 700 million people, the majority of whom are young and digitally connected, with internet penetration more than doubling over the past decade to over 73 percent in 2024, and the region's digital economy projected to approach $1 trillion by 2030. Yet despite this rapid digitalization and the region's strategic geopolitical positioning, Southeast Asia remains significantly underrepresented in global AI safety governance discourse, which continues to be shaped predominantly by the United States, China, and the European Union. Of the 10 ASEAN member states, six countries, Singapore, Indonesia, Malaysia, Thailand, Vietnam, and the Philippines, have achieved comparatively mature AI governance policy trajectories, while Laos, Myanmar, Brunei, Cambodia, and Timor-Leste lag behind in regulatory development and institutional capacity. Singapore stands out as the most developed ecosystem, ranked 11th globally and second in Asia and Oceania by the Global Index on Responsible AI (2024), while Laos and Myanmar rank in the bottom 15 percent of the same index.

The report identifies four key structural characteristics shaping the regional state of play: a universal recognition of AI's significance accompanied by nuanced national differences in perspective; a pragmatic, hindsight-informed approach to policy development that seeks to balance pro-business innovation with safety imperatives; diversity functioning as both a source of potential synergy and a barrier to regional regulatory harmonization; and persistent shared challenges, including a lack of quality datasets in low-resource languages, poor cybersecurity infrastructure, constrained governance capacity, and fragmented regulatory frameworks. These factors collectively underscore the urgency of inclusive, locally-grounded AI governance frameworks that can serve both national needs and contribute to the integrity of global standard-setting processes.

Brookings Report: Governing AI in the Southeast Asian Way

The report presents 10 actionable policy directives for the 2025–2030 period, addressed to both ASEAN-level institutions and national governments. At the regional level, the emphasis is placed on harmonizing governance approaches rather than enforcing regulatory uniformity, strengthening intraregional cooperation in capacity building, research, and catastrophic risk management, and engaging great-power dynamics with pragmatism. At the national level, the report calls for practical, high-impact measures including improvements in policy implementation, enhanced data collection infrastructure, proactive engagement with frontier AI risks, and streamlined institutional processes. Importantly, the report situates these directives within a broader call for more consistent and substantive Southeast Asian participation in international AI governance forums, noting that limited participation often reflects constrained administrative capacity rather than a lack of awareness or political will.

The report defines what it terms "AI safety governance, the Southeast Asian way", a model characterized by localized and inclusive governance, intraregional cooperation, and open-source safety tools. Key features include a focus on multilingual large language models (LLMs) adapted to the region's linguistic diversity, culturally aligned open-source safety toolkits, and ASEAN-led coordination mechanisms such as the Working Group on AI Governance. This approach offers a timely reference point for other Global Majority regions navigating the challenge of building AI governance frameworks that are both globally informed and locally responsive. As global AI governance norms harden and advanced capabilities advance toward artificial general intelligence, the report argues, Southeast Asia's voice must be amplified before inclusive policy windows close.

References

¹ Chua, Lyantoniette, Philip Tham, and Chinasa T. Okolo. "AI Safety Governance, the Southeast Asian Way." AI Safety Asia and Brookings Institution Center for Technology Innovation. August 2025.

² Association of Southeast Asian Nations. "ASEAN Guide on AI Governance and Ethics." ASEAN. 2024.

³ Association of Southeast Asian Nations. "Expanded ASEAN Guide on AI Governance and Ethics–Generative AI." ASEAN. January 2025.

⁴ Association of Southeast Asian Nations. "ASEAN Responsible AI Roadmap (2025–2030)." ASEAN. March 2025.

⁵ Global Index on Responsible AI. "Global Index on Responsible AI 2024 1st Edition Report." 2024.

⁶ Council of Presidents of the United Nations General Assembly (UNCPGA). "Governance of the Transition to Artificial General Intelligence (AGI): Urgent Considerations for the UN General Assembly." May 2025.

⁷ Carnegie Endowment for International Peace. "Speaking in Code: Contextualizing Large Language Models in Southeast Asia." January 2025.

⁸ World Economic Forum. "How ASEAN Is Building Trust in Its Digital Economy." January 2024.