Second co-creation workshop on the OECD AI Policy Toolkit and dialogue on "AI governance: Global issues, local approaches"

In the opening session, high-level speakers from Thailand, Japan, the United Kingdom and the OECD framed the workshop as a continuation of a longer term, cooperative effort to translate AI principles into practice. 

• H.E. Masato Otaka, Ambassador of Japan to Thailand, stressed the importance of sustained regional dialogue and internationally aligned approaches to safe, secure and trustworthy AI. 
• Gita Kothari, Director for Legal Affairs at the OECD, underlined that while global AI principles are increasingly well established, governments now face the more complex task of operationalising them through concrete policies, institutions and tools. 
• David Thomas, Deputy Head of Mission at the British Embassy in Thailand, highlighted the value of practical, experience based governance approaches that support innovation while maintaining public trust. 
• Prangtip Chansomsak, Deputy Director General at Thailand’s Ministry of Foreign Affairs, reaffirmed Thailand’s commitment to strengthening national AI governance while actively contributing to regional and global co-operation.

A dedicated session examined the Hiroshima AI Process (HAIP), a voluntary international framework aimed at strengthening transparency, risk management and learning in the governance of advanced AI systems, and experience to date with the HAIP Reporting Framework co-ordinated by the OECD – including private sector perspectives. 

The dialogue on AI safety drew on the International AI Safety Report 2026, including a presentation by one of its lead authors that summarised evidence on advances in general purpose AI capabilities, documented real world harms and the limitations of existing safeguards and evaluation practices. The session also benefited from a presentation outlining Singapore’s approach to managing AI risks through a combination of outcome based legislation, guidance and assurance mechanisms, including AI Verify.

A recurring theme was the importance of local context: AI systems often perform unevenly across languages and regions, which has implications for trust and effectiveness, particularly in public services. Several countries are therefore exploring more locally relevant approaches, including targeted language technologies and stronger testing frameworks, while recognising resource constraints and the importance of regional co-operation.

Discussions on AI compute highlighted the role of infrastructure as a key enabler of both AI development and effective governance. In that context, Thailand and Vietnam shared phased approaches to strengthening compute capacity, aligned with skills, energy and policy priorities. The OECD Secretariat, in turn, framed AI compute as a multi layered and distributed infrastructure requiring sustained strategic planning, and summarised policy trends as well as key findings from a recent OECD analysis of public cloud availability for AI. 

Discussions on AI in education and vocational training highlighted both opportunities and risks associated with AI deployment. The OECD Secretariat presented findings from the OECD Digital Education Outlook 2026, which highlighted both the potential of AI to improve learning outcomes when used appropriately and the risks of over reliance on automated tools. AI use in vocational education and training, including governance considerations related to data quality and accountability, was also addressed. 

In healthcare, discussions were anchored in the newly released OECD AI in Health Policy Checklist, which focuses on scaling AI while doing no harm. They emphasised the need to balance innovation with patient safety, equity and trust. Thailand showcased national efforts to use AI to for nutrition planning and preventative approaches (respectively, Thai School Lunch and KidDiary platforms), and Timor Leste shared plans for early stage use of AI supported disease surveillance and outbreak detection and response. 

Closing remarks were delivered by:

• Chai Wutiwiwatchai, Assistant Secretary of Thailand’s National AI Committee and Executive Director of NECTEC;
• Shige Watanabe, Deputy Assistant Minister for Economic Affairs, Ministry of Foreign Affairs of Japan;
• Emma Hardie, Head of AI International Engagement, UK Department for Science, Innovation and Technology
• Lucia Russo, Economist and Policy Analyst, OECD

In closing, speakers highlighted the value of the co creation approach and the strong interest in continued engagement. The upcoming OECD Dialogue on AI with Southeast Asian countries, supported by Japan, will contribute to moving forward on that ambitious yet crucial policy agenda.