Ministry of Post and Telecommunications

The first Dialogue is a success if: • It consolidates converging views on matters of concerns OR, otherwise, clearly captures the degrees and manners by which competing views differ. It must avoid becoming a place for broadcasting views, but no one listens to the others. Two key elements should be factored in: o The degree to which countries from the Global South can voice their genuine concerns and o The degree of authentic discussion with influential companies in the global AI market, where policymakers and executives interact meaningfully. • It offers a foundation to convert ideas into concrete actions, especially on matters of broad agreement. This requires the dialogue to enable participants to collectively prioritize issues and mobilize around concrete projects. Supply-demand (issue-solution) matchmaking initiatives should complement the formal dialogue. • It enables year-on-year progression. It should be possible for subsequent dialogues to build on the results of this first one. The dialogue's structure and mechanisms supporting it should aim to accumulate institutional memories within the UN and among member states and other participating entities. Working-level coordination between Dialogues is necessary. We may consider o Community of AI Governance comprising of working-level policymakers from member states and subject-matter experts from the private sector (not just government relations personnel) – with open directories of contacts. o Thematic discussions and activities throughout the year. The UN team needs to assess feasibility and desirability of each track before mobilization. Particular areas to begin: 1. Institutional capacity-building for regulators: technical expertise, regulatory /governance design, and support for enforcement mechanisms. 2. Technological and market intelligence: Enabling policymakers and regulators to better grasp emerging trends, risks, and supply chains market dynamics. Intelligence-sharing mechanisms would support evidence-based policymaking, enhance early risk detection, and reduce information asymmetries between developed and developing economies, empowering countries to engage more effectively in collective governance efforts.

• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Transparency, accountability, and human oversight
• Open-source software, open data and open AI models

The overarching issue of AI governance is that the technology develops so much more quickly than the mass population can adapt to it - a question of speed vs. scale. Thus, multi-pronged capacity building is the ultimate antidote to many potential issues. AI capacity-building includes training institutional leaders to make strategic decisions about AI adoption and risk management, increasing technical expertise particularly in developing countries, training people of different professions to respond to changes in their careers caused by AI, training governance and regulatory personnel, and training the mass population on AI literacy and staying relevant. Building on this, the full implications of AI technologies need to be continually studied and disseminated, so that stakeholders can make timely and proportionate measures, ensuring speedy adoption of AI and effective management of associated risks. The social and economic implications, particularly on the effects on citizens' wellbeing and income and national competitiveness are of paramount importance. The third priority - transparency, accountability, and human oversight - is a natural response to the nature of the technology and its current market dynamics, that is, powerful AI technologies are often inherently opaque, there's a critical imbalance in market powers, and trust needs to nurtured for both adoption promotion and inclusivity in managing this technology. Finally, open-source software, open data, and open models present a partial technology-based solution set to some of the emerging issues of AI. It is a strategic cross-cutting approach that contributes to capacity building, mitigating the economic imbalances caused by AI, and-consequently-creating more conducive conditions that support transparency, accountability, and human oversight with broad participation.

While the thematic areas identified by General Assembly Resolution 79/325 provide a strong foundation, several critical cross-cutting issues require explicit attention in the upcoming dialogue: • Environmental sustainability and climate impact of AI must be addressed. Exponential growth in computing power demands immense energy and water resources, necessitating governance frameworks that align AI development with global climate goals. • Invisible workforce behind AI development highlights the pressing need to address labor rights and the global AI supply chain. Data annotators, content moderators, and hardware manufacturers often face precarious working conditions, demanding the integration of fair labor standards across the entire AI lifecycle. • The issue of market concentration and geopolitical monopolies threatens equitable access. A handful of corporate entities and nations currently dominate AI infrastructure and foundation models. Addressing this concentration is vital to prevent the widening of the digital divide and to ensure that all states can meaningfully participate in and benefit from the global AI economy.

- Environmental sustainability and climate impact of AI: As AI adoption in Cambodia is still at an early stage, the country currently has limited capacity to assess or manage energy and resource implications of AI deployment, highlighting the need for sustainable digital infrastructure planning, especially when data centers and cloud services are rapidly expanding. This also presents an opportunity for Cambodia to integrate sustainable digital infrastructure planning and promote energy-efficient AI solutions, in line with the country's target of increasing renewable energy to around 70% of the energy mix by 2030. - Labor rights and the global AI supply chain: Cambodia currently faces limited institutional capacity to monitor or engage with international digital platforms on labor standards, making it difficult to ensure accountability in emerging forms of AI-related work. Many workers participate through informal, platform-based or cross-border arrangements without formal contracts, social protection, or access to grievance mechanisms, exposing them to job insecurity, low and inconsistent income, and, in some cases, harmful or sensitive content. At the same time, this presents an opportunity for Cambodia to strengthen governance frameworks for digital labor, enhance workforce protections, and invest in skills development, while leveraging international cooperation and knowledge-sharing to align with emerging global standards. - Market concentration and geopolitical monopolies: The concentration of advanced AI technologies and infrastructure in a limited number of global actors constrains Cambodia's access, increases dependency, and limits the development of locally relevant solutions. This challenge also presents an opportunity to promote open-source tools, strategic partnerships, and regional collaboration to strengthen local innovation capacity.

This Dialogue can serve as a global coordination platform to align governance approaches, share knowledge, and strengthen international cooperation. Importantly, it provides a space for both public and private sectors to engage directly, enabling governments, industry, and other stakeholders to openly discuss governance challenges and negotiate practical compromises on pressing issues such as safety standards, innovation constraints, market regulation, and the risks posed by harmful AI-generated synthetic media. The Dialogue should also promote technological and market intelligence-sharing mechanisms, helping regulators better understand rapid AI advancements, emerging risks, supply chain dynamics, and market concentration trends. This would strengthen evidence-based policymaking and reduce global information asymmetries. Additionally, the Dialogue can connect existing initiatives and foster interoperability, ensuring coherent and inclusive governance frameworks across regions. By combining multi-stakeholder dialogue, capacity-building, and intelligence-sharing, it can play a transformative role in shaping balanced, inclusive, and actionable global AI governance. To achieve broad legitimacy, the dialogue must be science-based and evidence-based, avoid undue ideological influence, and enable frank and constructive discussions between policymakers and private sector decision-makers.

#NAME?

All stakeholders play critical roles: - Governments: policy frameworks and regulatory insights - Industry: technical expertise and risk mitigation tools - Academia: evidence-based research - Civil society: human rights and societal impact perspectives The Dialogue should adopt a multi-stakeholder, hybrid format, ensuring meaningful participation and agenda-setting power for all groups. It should also include regional advisory panels and structured knowledge-sharing sessions. Importantly, dedicated tracks on regulatory capacity-building and technological/market intelligence should be included to support practical implementation. Pre-dialogue precipitation of common grounds are necessary to make the actual Dialogue productive. This demands some level of coordination for subject-matter discussions between stakeholders (with scientific opinion provided by the Independent Panel).

#NAME?

Innovative engagement formats should move beyond traditional plenary discussions and enable more interactive, solution-oriented exchanges. One effective approach is to convene multi-stakeholder roundtable sessions, bringing together policymakers, industry leaders, academia, and civil society in smaller, focused groups to deliberate on specific governance challenges—such as transparency, accountability, or risks from generative AI – preferably well in advance of the main Dialogue. These roundtables should be designed to encourage open, practical discussions grounded in real-world use cases, rather than high-level principles. By narrowing discussions to targeted issues (e.g., misinformation, workforce impacts, or sector-specific risks), participants can jointly explore trade-offs, share implementation experiences, and co-develop actionable recommendations. This format also helps ensure that a wider range of regional perspectives—particularly from developing countries—are included in more balanced and context-specific discussions. Additionally, there should be clear agreements on specific work prioritization, and subsequent regular meetings, to allow for concrete actions to be implemented after the discussion.

Several examples of policies, practices, platforms or approaches that promote effective AI governance or offer concrete solutions to addressing its challenges, include: 1. ASEAN Guide on AI Governance and Ethics (2024): the first regional instrument for AI governance, which focuses on balancing innovation and risk mitigation. It includes a set of 7 guiding principles: 1. Transparency and Explainability 2. Fairness and Equity 3. Security and Safety 4. Human-centricity 5. Privacy and Data Governance 6. Accountability and Integrity 7. Robustness and Reliability 2. Expanded ASEAN Guide on AI Governance and Ethics - Generative AI (2025): Build upon the fundamental AI Governance principles of ASEAN mentioned above and provide guidelines for the safe, responsible, inclusive development and application of Generative AI throughout the region. It demonstrates ASEAN's dedication to striking a balance between the requirements of growth, innovation, and trust, particularly for emerging economies. 3. OECD AI Policy Observatory: It is a global platform that support evidence-based AI policymaking and international coordination by functioning as a central hub that tracks and analyzes AI policies, regulatory developments, national strategies from across jurisdictions, enabling government to benchmark their approaches and identify best practices. Its key values lie in promoting transparency, comparability and interoperability, particularly supporting developing countries understanding global trends and designing context-appropriate AI governance frameworks. 4. Works similar to St. Gallen's Endowment's mapping of the different methods of translating principles into actions should be explored. For example, the UN can help map potential or experimented technological and regulatory solutions in response to particular areas of risks (e.g., deepfake content on social media can be dealt with using watermark of different technological approaches and/or using content regulations rules, such as Singapore's Manipulation of Online Falsehood Act and Online Safety (Accountability and Relief) Act). Such mapping will facilitate concrete discussions rather than high-level ones.