Trusts Motion

A successful first Global Dialogue on AI Governance would result in clear, actionable foundations for coordinated global cooperation rather than abstract principles alone. First, it should establish a shared baseline understanding of AI risks and opportunities, including agreement on what constitutes safe, responsible, and rights-respecting AI across different jurisdictions. This common grounding is essential for reducing fragmentation. Second, it should produce practical commitments toward interoperable governance frameworks, allowing countries to regulate AI in ways that are compatible rather than conflicting, while still respecting national sovereignty. Third, it should advance concrete mechanisms for international collaboration, such as shared safety evaluation standards, incident reporting systems, and joint research on AI safety and alignment. Fourth, it should ensure that capacity-building for developing countries is treated as a core outcome, not an optional add-on, so that all regions can meaningfully participate in AI development, oversight, and benefit-sharing. Finally, the dialogue should strengthen accountability structures by encouraging transparency norms, human oversight principles, and mechanisms for monitoring high-risk AI systems. A successful outcome would be a roadmap with timelines, responsibilities, and follow-up processes that keep momentum beyond the initial dialogue.

• Safe, secure and trustworthy AI
• AI capacity-building
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

These four priorities reflect the most urgent and foundational elements needed for responsible AI governance. Safe, secure and trustworthy AI is essential because without safety guarantees, other benefits of AI cannot be reliably realized. This includes reducing risks from misuse, unintended behavior, and systemic vulnerabilities. AI capacity-building is critical to ensure equitable participation. Without it, there is a growing divide where only a small number of actors shape AI systems, governance, and economic benefits. Strengthening skills, infrastructure, and institutional readiness is key for inclusivity. Protection and promotion of human rights is central because AI systems increasingly influence decisions affecting dignity, privacy, freedom of expression, access to services, and non-discrimination. Governance must ensure that technological progress does not undermine fundamental rights. Transparency, accountability, and human oversight are necessary to maintain public trust and enable effective regulation. AI systems must be explainable where appropriate, auditable, and subject to meaningful human control, especially in high-stakes contexts. Together, these priorities create a balanced framework that addresses safety, fairness, participation, and governance integrity.

Yes, several emerging issues are becoming increasingly important and are not fully captured by the listed themes. One key area is the governance of autonomous and agentic AI systems, where models can independently plan and execute tasks across digital environments. This raises new challenges around delegation of authority, liability, and real-time oversight. Another emerging issue is the concentration of compute, data, and AI capability within a small number of organizations and regions. This creates structural power imbalances that affect global equity and innovation diversity. Environmental impact is also increasingly important, particularly the energy and water consumption of large-scale AI training and deployment. Sustainability should become a core governance consideration. In addition, geopolitical fragmentation of AI standards and regulations could lead to incompatible ecosystems, reducing interoperability and increasing global risk. Coordinated mechanisms for alignment are needed to prevent regulatory isolation. Finally, there is a growing need for robust evaluation and auditing systems for AI models, including standardized benchmarking for safety, bias, reliability, and societal impact. Current evaluation methods are not yet mature enough for rapidly evolving systems. These cross-cutting issues highlight the need for adaptive, forward-looking governance that evolves alongside technological progress rather than reacting after harms occur.

Governance gaps in AI are increasingly shaping outcomes across public services, education, the digital economy, and institutional trust, with both risks and opportunities becoming more visible. One of the most significant challenges is the uneven capacity to regulate, deploy, and audit AI systems. In many contexts, including developing digital ecosystems, governance frameworks are still evolving, while AI technologies are advancing rapidly. This creates a lag where systems are already influencing decision-making in education, recruitment, public communication, and service delivery without sufficient oversight mechanisms in place. A related gap is the limited access to high-quality data infrastructure and AI expertise. This restricts the ability of institutions to independently evaluate or localize AI systems, increasing dependency on external providers. As a result, there is a risk of misalignment between imported AI models and local social, cultural, and linguistic contexts. There are also growing concerns around transparency and accountability. Many AI systems operate as "black boxes," making it difficult for public institutions and users to understand how decisions are made or to contest outcomes. This affects trust in digital services and can amplify perceptions of unfairness or exclusion. At the same time, there are significant opportunities. AI has the potential to improve efficiency in service delivery, expand access to education and mental health support, and strengthen data-driven policy design. In sectors such as youth development and mental health services, AI-enabled tools can help scale limited human resources and improve early intervention systems. Overall, the impact of governance gaps is a dual reality: they introduce risks of inequality and opacity, but also highlight the urgency and value of building inclusive, context-aware, and capacity-driven AI governance systems that enable safe innovation.

The AI Dialogue can serve as a neutral, inclusive platform that bridges fragmented global efforts on AI governance and helps move from parallel discussions to coordinated action. Its primary role should be to strengthen trust and shared understanding among states and stakeholders by creating a structured space for continuous exchange on risks, standards, and governance approaches. This is especially important in a context where AI development is uneven and regulatory maturity differs widely across countries. Second, the Dialogue can help identify minimum global baseline principles for safe and responsible AI, while still allowing flexibility for national implementation. This balance is critical to avoid regulatory fragmentation while respecting sovereignty. Third, it can facilitate practical cooperation mechanisms, such as shared safety evaluation methodologies, cross-border incident reporting frameworks, and common terminology for AI risk classification. These operational elements are essential for moving beyond high-level principles. Fourth, the Dialogue can play a catalytic role in supporting capacity-building partnerships, ensuring that developing countries are not left behind in both governance design and technical capability. Finally, it can act as a coordination hub that links existing initiatives, reduces duplication, and aligns efforts across international organizations, research institutions, and the private sector. In doing so, it would help transform AI governance from a scattered set of efforts into a more coherent global system.

The AI Dialogue should build upon and connect with several existing initiatives to avoid duplication and strengthen global coherence. Key foundations include UNESCO's Recommendation on the Ethics of Artificial Intelligence, the OECD AI Principles, and the G7 Hiroshima AI Process, all of which provide important normative frameworks. Additionally, technical and safety-focused efforts such as those led by the Global Partnership on AI (GPAI) and various national AI safety institutes contribute valuable research and evaluation practices. The Dialogue should also engage with standards development organizations working on AI interoperability and safety benchmarks, as well as UN-led digital cooperation mechanisms that already address broader issues of digital governance. Its added value would lie in creating a truly global, inclusive coordination platform that connects these fragmented efforts into a coherent ecosystem. Unlike many existing initiatives that are regional or thematic, the Dialogue can provide a universal space where developing and developed countries jointly shape governance priorities. It can also translate high-level principles into actionable governance pathways, including shared evaluation frameworks, risk classification systems, and capacity-building roadmaps. Importantly, it can serve as a bridge between technical AI governance communities and policy-makers, ensuring that governance frameworks are both technically grounded and socially legitimate. In this way, the AI Dialogue would not replace existing efforts but amplify their impact by improving coordination, inclusivity, and implementation consistency at the global level.

Different stakeholders should contribute through a structured, multi-layered participation model that ensures both technical depth and inclusive representation. Governments can provide policy direction, regulatory experiences, and national priorities, helping align global discussions with real implementation challenges. International organizations can facilitate coordination, neutrality, and continuity of the process. The private sector should contribute technical expertise, deployment insights, and information on emerging capabilities and risks, particularly regarding frontier AI systems. However, this should be balanced with transparency and accountability mechanisms. Academia and research institutions can provide independent analysis, evaluation frameworks, and evidence-based assessments of AI impacts across sectors. Civil society organizations should play a central role in highlighting societal, ethical, and human rights implications, ensuring that governance remains people-centered. To structure participation effectively, the Dialogue should adopt a tiered format combining: High-level plenaries for political and strategic alignment Technical working groups focused on specific governance themes Open multi-stakeholder forums for broader input and transparency Regional consultations to capture local and contextual realities This structure ensures that discussions are both globally coherent and locally grounded, while allowing sustained engagement beyond annual meetings.

Several important voices remain underrepresented in global AI governance discussions. First, developing countries and smaller economies often lack equal participation due to resource and capacity constraints, despite being significantly affected by AI deployment. Their inclusion requires targeted funding, capacity-building programs, and equitable representation in decision-making spaces. Second, youth and future generations are rarely directly represented, even though AI will shape their long-term social and economic environment. Structured youth panels and advisory groups can help integrate their perspectives. Third, local and indigenous communities are often excluded, despite their knowledge systems and cultural contexts being highly relevant to ethical and inclusive AI design. Engagement must be culturally sensitive and community-led. Fourth, workers in informal sectors and those affected by AI-driven automation are often missing from policy discussions. Their lived experience is essential for understanding socio-economic impacts. Finally, small research institutions and independent researchers outside major tech hubs face barriers to participation due to funding and visibility constraints. Inclusion can be improved through hybrid participation models (in-person and virtual), multilingual access, travel and participation support funds, and structured consultation mechanisms that actively seek out underrepresented groups rather than relying on open calls alone.

To foster meaningful engagement, the AI Dialogue should move beyond traditional conference formats and adopt more interactive, continuous, and collaborative models. One effective approach is the use of "living labs" or simulation environments where stakeholders can test governance scenarios, such as AI deployment in healthcare, education, or public services. This allows participants to experience real-world trade-offs rather than discussing them abstractly. Another innovative format is structured "policy hackathons," where multidisciplinary teams collaboratively design governance solutions for specific AI challenges within a limited timeframe. The Dialogue could also include interactive digital platforms that enable continuous participation before, during, and after formal sessions. These platforms could host deliberation forums, voting on proposals, and collaborative drafting of policy frameworks. Scenario-based foresight workshops are also valuable, enabling stakeholders to explore future risks and opportunities under different AI development trajectories. Additionally, rotating regional dialogue hubs could decentralize participation and ensure that discussions are grounded in diverse socio-economic contexts. Finally, pairing technical experts with non-technical stakeholders in "tandem panels" can improve mutual understanding and reduce the gap between policy language and technical realities. Together, these formats would make the Dialogue more participatory, practical, and adaptive, ensuring that outcomes are not only discussed but actively co-created.

Several emerging and established approaches demonstrate practical pathways toward effective AI governance and can inform global coordination efforts. A strong example is the OECD AI Principles, which provide a widely adopted baseline for trustworthy AI, emphasizing robustness, transparency, human-centered values, and accountability. Their strength lies in their adaptability across different legal and cultural contexts. The UNESCO Recommendation on the Ethics of Artificial Intelligence is another important framework, particularly for its emphasis on human rights, inclusion, and sustainability. It provides actionable guidance for national policy development, especially in education, culture, and public sector deployment. At the regional level, the European Union AI Act represents one of the most comprehensive regulatory frameworks, introducing a risk-based approach that categorizes AI systems by level of harm and imposes proportionate obligations. While still evolving, it offers a structured model for balancing innovation with safeguards. In addition, the Global Partnership on AI (GPAI) contributes valuable multi-stakeholder research and practical experimentation, particularly in areas such as responsible AI deployment, data governance, and algorithmic accountability. Nationally, several countries have established AI safety institutes and regulatory sandboxes, allowing for controlled testing of AI systems before full-scale deployment. These mechanisms help bridge the gap between innovation and regulation. In the private sector, emerging model evaluation and auditing frameworks, including red-teaming practices and transparency reporting, are becoming important governance tools that enhance accountability and system safety. Collectively, these approaches demonstrate that effective AI governance is most successful when it combines risk-based regulation, international cooperation, technical standards, and inclusive multi-stakeholder participation.