AI 2030

Success for the first Global Dialogue on AI Governance should be measured across three levels: structural, practical, and human. Structurally, success means a formalized multi-stakeholder coalition with representation from at least 50 countries across all UN regional groups, including meaningful participation from Global South institutions as co-architects of governance frameworks. A governance charter and working group structure should be established within 90 days of the dialogue's close. Practically, success means delivery of at least 3 to 5 co-designed, globally deployable governance toolkits developed across civil society, academia, private sector, and government. These tools should be adaptable across jurisdictions and usable without requiring mature regulatory infrastructure. Pilot implementation should occur in at least 10 countries representing diverse governance capacity levels within 12 months. At the human level, success means measurable progress toward protection at the point of AI use, not only institutional compliance. This includes adoption of individual-facing frameworks, such as metacognitive and self-regulatory tools, by at least 25 organizations across education, workforce development, and civil society sectors globally. Baseline indicators should be established to assess whether governance reaches individuals, with data disaggregated by region, gender, and development context. Finally, success requires coordination, not duplication. The dialogue should map linkages to existing fora such as IGF, OHCHR, OECD, and WSIS, with clear accountability and a public progress report within 6 months. The measure of success is not the quality of the document produced, but whether the person in Kumasi, Lagos, or Bogotá has the tools, agency, and liberty to engage AI safely one year from now.

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• AI capacity-building

These priorities reflect the need to connect governance with real-world human interaction with AI systems. Safe, secure, and trustworthy AI depends not only on system design, but on how individuals engage with AI in practice. Individuals must be equipped to assess intent, evaluate risk, and make informed decisions in real time. AI capacity-building is essential at every level, from institutions to everyday users. This includes co-designed, open-source tools and learning pathways that are globally accessible, particularly in contexts with limited governance infrastructure. Open-source approaches expand reach, reduce cost barriers, and strengthen local teaching capacity where it is needed most. Protection and promotion of human rights requires preserving individual agency in AI-mediated environments. This includes the ability to engage AI safely, recognize potential harms, and act with informed awareness. Social, economic, ethical, cultural, linguistic, and technical implications of AI must be addressed through human-centered and contextually responsive frameworks, not designed for one governance context and exported to others. Together, these priorities point to the need for practical tools and shared frameworks that support responsible AI use at the point of interaction. Without this integration, governance risks remaining abstract while individuals navigate AI systems without sufficient support.

A critical cross-cutting issue not addressed by the listed themes is human protection at the point of AI interaction. Current frameworks emphasize system-level governance, institutional accountability, and policy alignment. However, individuals are already engaging with AI systems in real time, often without practical tools to assess intent, evaluate risk, or respond appropriately in the moment of use. This creates a gap between governance and lived experience. Addressing this gap requires approaches that support individuals directly, including metacognitive and self-regulatory frameworks that can be applied across contexts and levels of governance maturity. A related issue is the need to operationalize co-design as a standard practice. This includes engaging diverse stakeholders through empathetic listening and collaborative development processes to ensure that tools, frameworks, and policies are usable, relevant, and globally adaptable. In addition, greater attention is needed on how governance is experienced by individuals, not only how it is defined at the institutional level. This includes developing indicators that measure whether people have the awareness, capacity, and tools to engage AI safely and with agency. These issues cut across safety, human rights, transparency, and capacity-building. Without addressing them, there is a risk that governance remains abstract while individuals continue to navigate AI systems without sufficient support.

Across sectors and regions, a key governance gap is the disconnect between rapidly advancing AI capabilities and the practical ability of individuals and institutions to engage these systems safely and responsibly. In many contexts, policy development is progressing, but implementation remains uneven. This is particularly evident in regions with emerging governance capacity, where institutions may lack the infrastructure, training, or resources to translate principles into practice. At the same time, individuals are already interacting with AI systems daily, often without clear guidance or tools to assess risk, recognize manipulation, or make informed decisions in real time. This creates both a challenge and an opportunity. The challenge is that governance can remain abstract, with strong frameworks at the institutional level but limited impact at the point of use. This increases the risk of misuse, misunderstanding, and uneven outcomes across regions and populations. The opportunity is to bridge this gap through practical, human-centered approaches. This includes co-designed tools, open and accessible learning pathways, and frameworks that support individuals in assessing intent, evaluating risk, and engaging AI responsibly in real-world contexts. In the education and workforce sectors, there is significant potential to embed these approaches into training, credentialing, and lifelong learning systems. This would strengthen both individual capacity and institutional readiness. Addressing these gaps requires aligning governance, capacity-building, and human-centered design to ensure that AI is not only well-regulated, but also safely and effectively used in practice.

The AI Dialogue can serve as a bridge between governance and real-world use by advancing shared, co-designed frameworks that support responsible AI at the point of interaction. Its role is not only to align principles, but to enable practical application through tools that help individuals assess intent, evaluate risk, and act with informed awareness in real time. A core function is fostering co-ownership across stakeholders so responsibility does not sit only with developers or policymakers, but is shared by those who design, deploy, and use these systems. The Dialogue can also elevate insights from frontline users and translate them into globally adaptable approaches grounded in empathetic listening and co-design. By centering co-ownership and felt accountability, the AI Dialogue can ensure that international cooperation produces responsible action, not just agreement.

The AI Dialogue should build on existing governance frameworks, standards efforts, and workforce and education initiatives already shaping responsible AI. Its added value is the ability to connect these efforts through shared, co-designed frameworks that support use at the point of interaction. Rather than duplicating existing work, the Dialogue can: →•Align frameworks across sectors to create consistency in how responsible AI is understood and applied →•Support the development of practical toolkits that accompany these frameworks →•Enable global deployment through adaptable models that reflect local context and capacity A key contribution is strengthening the link between systems and users by ensuring that frameworks are not only defined, but actively used. By focusing on co-design, shared frameworks, and practical deployment, the Dialogue can help move from fragmented efforts to coordinated, actionable practice.

Different stakeholders should contribute through co-designed, role-based participation that reflects how they engage with AI in practice. The Dialogue should be structured to include policymakers, developers, educators, employers, and everyday users, each contributing from their point of interaction with AI systems. This can be supported through small working groups, case-based discussions, and structured input formats that prioritize real-world experience. The format should emphasize empathetic listening and co-design, allowing stakeholders to surface challenges, test ideas, and contribute to shared frameworks and tools. Clear pathways for contribution should be paired with feedback loops so participants can see how their input informs outcomes. This structure supports co-ownership and strengthens felt accountability across the ecosystem.

Voices that are often underrepresented include frontline workers, educators, learners, small and medium enterprises, communities in low-resource settings, and individuals interacting with AI outside formal governance structures. These groups experience AI directly but are often excluded from shaping how it is governed. Inclusion can be strengthened by creating accessible participation pathways, including simplified input formats, multilingual access, and partnerships with local organizations that can convene and represent community perspectives. Co-designed tools and frameworks should reflect diverse cultural, linguistic, and economic contexts rather than assuming a single model of governance. Ensuring that these voices are not only included but reflected in outcomes strengthens both relevance and trust.

Effective engagement should move beyond static consultation toward interactive, practice-based formats. This can include: •Scenario-based discussions grounded in real AI use cases •Co-design sessions to develop shared frameworks and tools •Small, facilitated groups that encourage participation and dialogue •Digital platforms that allow ongoing input, not just one-time responses Blending in-person and virtual formats can expand access while maintaining depth of engagement. The most effective formats will center participation, enable contribution from diverse stakeholders, and create visible connections between input and action. This approach reinforces co-ownership and supports meaningful, sustained engagement.

One effective approach is the use of metacognitive frameworks to support responsible AI use at the point of interaction. The MARIT framework, Mindful AI and Regulation in Technology Use, equips individuals to assess intent, evaluate risk, and engage AI responsibly in real time. It has been applied across military training, higher education, and international contexts, including NATO and the Ukrainian Defence University, demonstrating adaptability across governance environments. A complementary approach is embedding AI governance into credentialing and workforce development systems. The AI 2030 Institute CRAFT initiative, Credentialing Responsible AI for Future-Ready Talent, is a national benchmark effort designed to establish shared standards for responsible AI competency across education and employer sectors. This approach connects governance principles to measurable workforce outcomes. Together, these approaches illustrate how governance can be operationalized at the individual and institutional level, not only defined at the policy level. Key design principles include co-design with diverse stakeholders, open and accessible formats that reduce barriers to adoption, and alignment with existing learning and credentialing infrastructure. These models are globally adaptable and can be deployed across contexts with varying levels of governance maturity, making them particularly relevant where capacity-building and practical implementation are urgent priorities.