Southeast Asian governments are actively exploring frameworks for effective artificial intelligence (AI) governance as part of a regional effort to maximize the technology’s potential while managing its risks.
Over the past year, the Organisation for Economic Co-operation and Development (OECD), with backing from Japan and the United Kingdom, has engaged with Southeast Asian countries through a series of dialogues and workshops. Notable events have included the OECD AI Policy Roundtable in Tokyo, Japan, co-creation workshops on the OECD AI Policy Toolkit in Bangkok, and the OECD-Southeast Asia Policy Roundtable focusing on sectors like shipbuilding and steel.
These dialogues reveal a consensus that the primary challenge has shifted from establishing AI governance principles to effectively implementing them in practice. Participants are eager to move forward with concrete actions to enhance policy development and execution.
Implementing the OECD AI Policy Toolkit
The OECD AI Policy Toolkit serves as a resource to help governments navigate their AI policy landscapes, prioritize actions, and draw upon international best practices. The toolkit is designed for domestic policy efforts and encourages peer learning rather than ranking countries based on AI performance.
During discussions in Bangkok, participants expressed a strong desire to translate principles into practice. Specifically, they sought guidance on how the toolkit could assist in identifying policy development priorities, foster coordination among government ministries, and be customized for diverse national contexts.
Feedback indicated that countries aim not only to recognize gaps but also to prioritize and sequence their implementation of AI policies, with participants valuing the toolkit’s potential to inform decision-making and policy development.
Addressing AI challenges and uncertainties
As AI technologies evolve, governments face the challenge of making informed decisions amid emerging evidence about their impacts. The International AI Safety Report 2026 highlights significant advancements in AI alongside risks such as fraud and cybersecurity threats. Additionally, safeguards may not always yield the expected outcomes in uncontrolled environments.
In response, some nations are adopting practical strategies. For instance, Singapore is leveraging existing legislation while incorporating voluntary guidelines like AI Verify to facilitate ongoing testing and facilitate adaptive regulatory approaches.
Collaborative frameworks for AI governance
Effective AI governance requires participation from stakeholders beyond government entities. The Hiroshima AI Process (HAIP) Reporting Framework invites organizations involved in developing advanced AI systems to disclose their risk management practices. This framework aims to create transparency and consistency in AI operations and contribute to a better understanding of AI risks.
Going forward, three keys to successful AI governance will be expanding stakeholder participation, keeping pace with rapid technological advances, and promoting interoperability among emerging AI governance frameworks.
Cultural and linguistic challenges
A notable concern in AI application is the uneven performance of systems across different languages. Inaccurate translations or lost context can affect the effectiveness of AI, particularly in regions with diverse language needs. This issue has prompted initiatives focused on developing AI models better suited to Southeast Asian languages and cultural context, with countries like Japan sharing resources for data and model development.
Despite progress, constraints on high-quality training data for local languages hinder the development of effective AI systems and the ability to tailor large global models to regional needs. Singapore is pursuing multilingual testing strategies to ensure AI systems work reliably in their local contexts.
Investment in AI infrastructure
AI compute capacity is increasingly being viewed as critical infrastructure that encompasses skills and cloud service access. In Bangkok discussions, Thailand highlighted an approach linking AI investments with public sector usage and skills training, while Vietnam outlined a more gradual strategy focused on model training and evaluation.
As countries seek to build their AI capabilities, access to compute resources will be vital for innovation and effective governance, enabling nations to accurately assess and manage AI systems.
Practical applications in education and healthcare
Recent sessions on education and healthcare highlighted the complexities of embedding AI within existing frameworks. The OECD shared insights from its Digital Education Outlook 2026, emphasizing that while AI can enhance learning, it requires clear pedagogical frameworks to avoid negative outcomes. Meanwhile, Thailand showcased projects like Thai School Lunch and KidDiary, which use AI for nutrition planning and child development monitoring, respectively.
These examples demonstrate that the success of AI initiatives is shaped by how well they integrate into existing governance and service delivery systems.
Strengthening regional AI cooperation
The OECD’s engagement with Southeast Asia is gaining traction with recent workshops focusing on harnessing responsible AI. Following the inaugural OECD-Southeast Asia AI Dialogue in September 2026, a second workshop took place in November to promote regular exchanges among countries on challenges like risk management and infrastructure needs. The next gathering is scheduled for Indonesia, aiming to deepen collaboration on shared AI objectives.
As countries continue to refine their approaches and learn from one another, the foundational principles of AI governance are being adapted to local contexts, ensuring that technology serves the needs of Southeast Asian societies effectively.


