AVSE Global

A successful first Global Dialogue on AI Governance should move beyond discussion and deliver tangible, action-oriented outcomes that can guide global cooperation. First, the Dialogue should result in a shared global framework or code of conduct on AI ethics. While many principles already exist, there is fragmentation across regions and institutions. A UN-facilitated, globally endorsed set of ethical standards would provide a common reference point to guide governments, industry, and civil society, particularly ensuring that AI systems are safe, transparent, accountable, and human-centered. Second, success would mean establishing inclusive governance mechanisms that meaningfully integrate voices from the Global South. AI governance must not be dominated by a few technologically advanced countries; instead, it should reflect diverse realities, especially from regions where regulatory capacity and digital infrastructure are still evolving. Third, the Dialogue should catalyze practical cooperation on capacity-building, particularly in education and workforce readiness. This includes advancing AI literacy, digital skills, and ethical awareness among young people to ensure that societies are not only protected from AI risks but are also empowered to benefit from its opportunities. Finally, a key outcome would be the creation of clear follow-up pathways, including multi-stakeholder working groups, knowledge-sharing platforms, and measurable commitments leading into the 2027 Dialogue. Ultimately, success lies in ensuring that AI governance becomes not just a global conversation, but a shared global responsibility with concrete direction, inclusive participation, and real-world impact.

• Transparency, accountability, and human oversight
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Safe, secure and trustworthy AI
• Interoperability of governance approaches

My selection of these four themes reflects both their interconnected nature and their critical relevance to ensuring that AI governance is not only technically sound, but also socially just and globally inclusive. First, safe, secure and trustworthy AI is foundational. In many developing contexts, where regulatory systems are still evolving, trust in AI systems will determine the extent to which societies are willing-and able-to adopt them responsibly. Second, the theme of social, economic, ethical, cultural, linguistic and technical implications is essential to ensure that AI does not deepen existing inequalities. From my experience working with youth and higher education systems in Vietnam and Southeast Asia, I have seen how gaps in digital access, language representation, and institutional capacity can limit who benefits from technological advancement. Third, interoperability of governance approaches is critical in a fragmented global landscape. Without some level of alignment, differing national regulations may create barriers to cooperation, innovation, and equitable participation, particularly for countries with more limited resources. Fourth, transparency, accountability, and human oversight are key to operationalizing AI governance. Principles alone are insufficient; what matters is how they are implemented through concrete mechanisms that ensure responsibility and public trust. Together, these themes reflect a comprehensive approach, one that balances innovation with responsibility, and global ambition with local realities. They are particularly important to ensure that AI governance frameworks remain inclusive, actionable, and responsive to the needs of diverse societies.

A critical cross-cutting issue not fully captured in the current themes is the AI-climate nexus, which presents both strategic opportunities and systemic risks. AI has the potential to accelerate climate action, enhancing early warning systems, climate modeling, and resource optimization, particularly in highly vulnerable regions. For many developing countries, this could significantly strengthen resilience and adaptive capacity. However, this opportunity is coupled with a growing and often overlooked challenge: the environmental cost of AI itself. The rapid scaling of AI models and data infrastructure is driving increased energy consumption and carbon emissions. Without clear governance, AI risks becoming both a solution to-and a contributor of-the climate crisis. This dual impact underscores a critical governance gap. AI governance must explicitly integrate environmental sustainability as a core principle, not a peripheral concern. This includes: • Standardized measurement and disclosure of AI-related emissions • Incentives for energy-efficient and low-carbon AI development • Alignment between AI strategies and national and global climate commitments Positioning climate as a cross-cutting issue will ensure that AI development is not only innovative, but also ecologically responsible and future-proof.

In Vietnam and across Southeast Asia, governance gaps in AI are creating a high-stakes dual dynamic, accelerating opportunity while amplifying systemic risks. Key challenges stem primarily from fragmented and underdeveloped governance frameworks. At the regional level, existing guidance, such as ASEAN's AI governance principles, remains non-binding and inconsistently implemented, limiting enforcement and accountability. Nationally, countries like Vietnam are advancing AI adoption in public services, yet legal frameworks, data governance standards, and AI ethics regulations remain incomplete. This is compounded by: - Uneven readiness and capacity gaps across countries and institutions, making harmonized governance difficult - Limited technical expertise within government, constraining effective oversight - Risks related to data privacy, algorithmic bias, and misinformation, which current systems are not fully equipped to manage At a broader level, there is a growing concern that AI could widen existing inequalities, both within and between countries, if governance does not keep pace with technological advancement. However, these gaps also create significant opportunities. AI is emerging as a key driver of economic transformation in the region, with the potential to contribute up to nearly $1 trillion in GDP by 2030. Vietnam, with its strong digital transformation agenda and young workforce, is well-positioned to leverage AI to improve public services, education, and inclusive development.

The AI Dialogue can play a pivotal role as a global coordination platform to advance international cooperation on AI governance, particularly by enabling structured exchange, alignment, and shared accountability across countries and sectors. First, it can facilitate the systematic sharing of governance experiences and best practices. Countries are currently developing AI policies at different speeds and with varying approaches. The Dialogue can help bridge this fragmentation by creating a space where governments, industry, and civil society exchange lessons learned, including both successes and failures, to inform more adaptive and context-sensitive governance models. Second, the Dialogue can promote responsible data collaboration at the international level. Data is at the core of AI development, yet access remains uneven and governance standards are inconsistent. By encouraging cooperation on data-sharing frameworks, interoperability standards, and ethical data use, the Dialogue can help unlock innovation while safeguarding privacy, security, and national interests. Third, it can mobilize resources and capacity-building across sectors. Many countries, particularly in the Global South, face constraints in technical expertise, infrastructure, and regulatory capacity. The Dialogue can act as a connector, linking governments with international organizations, academia, and the private sector to support knowledge transfer, funding, and technical assistance. Finally, the AI Dialogue can reinforce transparency and accountability as global norms. By encouraging voluntary commitments, reporting mechanisms, and multi-stakeholder oversight, it can help build trust and ensure that AI governance is implemented responsibly across borders. Ultimately, the Dialogue's value lies in transforming fragmented national efforts into coordinated, transparent, and inclusive global action on AI governance.

At the global level, the AI Dialogue should build upon existing United Nations–led frameworks, including the UN General Assembly resolution on AI (2024) and broader efforts under the Global Digital Compact. These initiatives have established important normative foundations around safe, inclusive, and human-centered AI, but remain largely non-binding. The AI Dialogue can add value by translating these high-level principles into actionable guidance, while strengthening coordination, transparency, and accountability across countries through a more structured and continuous multilateral process. At the regional level, the European Union has taken a leading role through the EU AI Act, which represents the most comprehensive, risk-based regulatory framework currently in place. Its legally binding nature and detailed compliance mechanisms offer valuable lessons for operationalizing AI governance. However, its applicability beyond Europe remains uneven. The AI Dialogue can help bridge such regional standards with global needs, facilitating interoperability and ensuring that regulatory approaches are adaptable to diverse legal, economic, and institutional contexts. In Southeast Asia, ASEAN has developed non-binding AI governance guidelines that emphasize inclusivity, innovation, and capacity-building. While these frameworks are well-aligned with development priorities, their implementation is constrained by varying levels of readiness across member states. The AI Dialogue can play a critical role in amplifying regional voices, supporting capacity-building, and connecting ASEAN frameworks with global standards, ensuring that developing countries are not only rule-takers but active contributors to shaping the future of AI governance.

Advancing effective AI governance requires meaningful contributions from diverse stakeholders, each bringing distinct expertise and responsibilities. The AI Dialogue should be structured to enable balanced, action-oriented participation across four key groups. Governments play a central role in shaping regulatory frameworks and ensuring alignment with public interest. They should contribute by sharing policy approaches, identifying governance gaps, and committing to transparent, interoperable standards. The scientific and research community provides the technical foundation for informed decision-making. Their role is critical in translating complex AI developments into evidence-based guidance, risk assessments, and practical tools that policymakers can apply. Private sector actors, particularly AI developers and technology companies, are essential for operationalizing governance. They should contribute through responsible innovation practices, transparency in model development, and participation in setting industry standards and accountability mechanisms. Users and civil society bring perspectives on real-world impacts, inclusion, and rights. Their engagement ensures that AI governance remains people-centered, reflecting societal needs and protecting vulnerable groups. To maximize impact, the AI Dialogue should adopt a multi-layered and outcome-driven structure. This could include: • Thematic working groups aligned with priority areas, producing concrete policy recommendations • Multi-stakeholder roundtables to foster cross-sector dialogue and consensus-building • A knowledge-sharing platform to exchange data, best practices, and case studies • Clear follow-up mechanisms, including voluntary commitments and progress reporting Such a structure would enable the Dialogue to move beyond discussion toward coordinated, transparent, and accountable global action on AI governance.

Global discussions on AI governance continue to be dominated by governments, large technology companies, and technical experts, while several critical voices remain underrepresented, particularly children and adolescents, persons with disabilities, and women, especially from developing countries. These groups are not only disproportionately affected by AI systems, but also face structural barriers to participation. For example, children and young people are among the most active users of digital technologies, yet their perspectives on safety, privacy, and digital well-being are rarely reflected in policy processes. Similarly, persons with disabilities risk exclusion if AI systems are not designed with accessibility in mind, while gender biases in data and algorithms continue to reinforce inequalities affecting women. To address this gap, AI governance must move beyond traditional consultation models toward more inclusive and participatory approaches. First, the AI Dialogue should promote open global forums and accessible consultation platforms, enabling diverse communities to contribute their perspectives in safe and meaningful ways. This includes using multiple languages, accessible formats, and youth-friendly engagement methods. Second, innovative mechanisms such as global idea challenges, youth forums, and community-led innovation competitions on AI can help amplify underrepresented voices and generate grassroots solutions. These platforms not only encourage participation, but also empower communities to shape how AI is designed and governed. Third, partnerships with civil society organizations are essential to ensure sustained engagement and representation of vulnerable groups in policy discussions. Ensuring inclusive participation is not only a matter of equity, it is fundamental to building AI systems that are fair, representative, and responsive to the needs of all.

To foster meaningful and dynamic engagement, the AI Dialogue should move beyond traditional plenary discussions toward interactive, participatory, and solution-oriented formats that enable real co-creation among stakeholders. First, multi-stakeholder co-creation labs could serve as core engagement spaces. Bringing together governments, researchers, private sector actors, and civil society in small, diverse groups would allow participants to collaboratively develop practical policy solutions, prototypes, or governance models around specific challenges. Second, the Dialogue could integrate global idea challenges and innovation competitions on AI governance, particularly targeting youth and underrepresented communities. These formats can surface fresh perspectives and grassroots solutions, while democratizing participation beyond formal policy actors. Third, scenario-based simulations and policy hackathons would enable participants to test governance responses to real-world AI risks such as misinformation, bias, or system failures, in a controlled, interactive setting. This approach strengthens both preparedness and cross-sector understanding. Fourth, the use of hybrid digital platforms is essential to ensure global inclusivity. Virtual participation tools, multilingual interfaces, and open-access knowledge hubs can expand engagement to stakeholders who cannot be physically present, particularly from the Global South. Finally, the Dialogue should incorporate commitment-driven sessions, where stakeholders publicly present voluntary actions, partnerships, or pilot initiatives, followed by transparent tracking mechanisms. By combining co-creation, competition, simulation, and digital inclusion, the AI Dialogue can evolve from a forum for discussion into a platform for innovation, collaboration, and measurable global action.

A notable example of effective AI governance is Vietnam's approach to aligning regulatory development with its broader national digital transformation strategy, positioning AI as a key driver for long-term socio-economic development. Through its National Strategy on AI to 2030, Vietnam frames AI as a strategic enabler for achieving its vision of becoming a high-income, developed country by 2045. This ensures that AI governance is embedded within national priorities on economic growth, public service innovation, and social inclusion, rather than treated as a standalone technical issue. Importantly, this strategic vision is being reinforced by emerging regulatory frameworks, including Vietnam's Law on Artificial Intelligence (2025). The law reflects a risk-based and forward-looking governance approach, aiming to balance innovation with safeguards related to safety, accountability, and ethical use of AI. It signals a transition from high-level principles toward more structured and enforceable governance mechanisms. A key strength of Vietnam's model lies in its policy coherence, linking national strategy, legal frameworks, and practical implementation. AI is being applied in areas such as public administration and digital services, while governance systems continue to evolve to address challenges in data management, transparency, and oversight. This integrated approach demonstrates that effective AI governance requires not only regulation but also a clear strategic alignment with national development goals. Vietnam's experience offers a valuable example of how countries can simultaneously advance innovation, governance, and inclusive development in the age of AI.