International Chamber of Commerce

The first Global Dialogue on AI Governance will be most valuable if it is understood as the beginning of a sustained, structured process rather than a standalone event. It would be successful if it delivers clear, practical outcomes that strengthen understanding of how AI is deployed in practice, while establishing a trusted and inclusive foundation for ongoing engagement. A successful first Dialogue would establish clear thematic priorities to guide subsequent work, rather than attempting to address all issues under its mandate at once; enable information sharing and produce actionable, practical outputs that enable the implementation of existing principles; and put in place complementary, light-touch mechanisms, such as intersessional working groups, to foster continuity and maintain momentum between annual sessions. The Co-Chairs' Summary will be critical in capturing these outcomes. A strong Summary should reflect practical, evidence-based, and action-oriented conclusions, capturing concrete solutions, areas of convergence and opportunities for collaboration. Another factor of success would be the Dialogue's ability to draw on practical input from stakeholders, including concrete case studies, evidence, and actionable recommendations that move beyond high-level principles. This will also depend on the Dialogue's ability to build on existing global networks and multistakeholder platforms to consolidate stakeholder input in an ongoing, structured way, e.g. informal information-sharing sessions and opportunities to contribute best practices grounded in real-world experience. This would help fill existing evidence gaps and ensure discussions complement, rather than duplicate, work already underway in other fora such as the OECD, the G20, or regional bodies. Finally, maintaining strong linkages with major global AI and digital policy convenings will also be important to ensure coherence across the global AI governance landscape and allow external developments to meaningfully inform the Dialogue.

• Safe, secure and trustworthy AI
• AI capacity-building
• Interoperability of governance approaches
• Transparency, accountability, and human oversight

Interoperability of governance frameworks is a foundational issue that underpins the effectiveness of global AI governance. In a global digital economy, AI systems, data flows and value chains operate seamlessly across borders, yet regulatory approaches risk becoming fragmented, duplicative or inconsistent. Divergent national or regional rules can undermine trust, raise compliance costs, stifle innovation, and disproportionately affect SMEs and actors in developing and least developed countries. Interoperable governance does not require uniform regulation, but rather alignment around common principles, risk-based approaches and shared terminology, enabling different legal systems to work together coherently. This is particularly critical for advancing safe and trustworthy AI, as fragmented standards for risk management, safety testing, or accountability can weaken protections rather than strengthen them. Likewise, transparency, accountability, and meaningful human oversight are more effective when governance approaches are mutually understandable and supported by internationally recognised best practices. Capacity building is also inseparable from interoperability. Without global coherence, capacity-building efforts risk preparing stakeholders to comply with a patchwork of incompatible rules, limiting participation from developing economies. Prioritising those issues will allow for the development of AI policy approaches that support trust, innovation and inclusion, while enabling businesses of all sizes to operate responsibly and competitively across markets.

Organizing the issues contained in the Dialogue's mandate into thematic groupings can make the overall programme more coherent and easier to navigate. A structured approach helps participants identify where discussions are taking place, understand how related topics connect, and engage more effectively across different parts of the Dialogue. Clustering issues, therefore, can support broader participation and improve the accessibility of the process. At the same time, the issue of interoperability and compatibility of governance approaches should remain a distinct, stand-alone priority, rather than incorporated within another grouping. Its relevance cuts across multiple areas of the Dialogue, but its importance lies specifically in addressing how policy and regulatory frameworks interact across jurisdictions. As AI governance frameworks continue to emerge globally, differences in regulatory approaches can create fragmentation, increase compliance complexity, and hinder international cooperation. A dedicated focus on interoperability would create space to examine how governance systems can remain compatible while reflecting different national priorities and legal traditions. This includes consideration of shared principles, regulatory coordination, mutual recognition mechanisms, and approaches that reduce unnecessary divergence. Positioning interoperability as an independent theme would strengthen the Dialogue's ability to address practical governance challenges at the international level. It would also reinforce the Dialogue's role in supporting greater coherence across evolving policy frameworks, helping ensure that global AI governance develops in a more interoperable manner.

From a private sector perspective, the most significant challenge arising from current governance gaps is regulatory fragmentation and the lack of interoperable governance approaches, including divergent approaches to data governance and cross border data flows, which can limit the ability of companies to scale AI solutions across markets. Divergent – and sometimes even conflicting – requirements related to AI risk classification, transparency, documentation, and conformity assessment can increase compliance costs and uncertainty, particularly for SMEs with limited legal and technical capacity. Fragmentation can also slow AI diffusion, as AI development and deployment depend on access to high quality data, compute infrastructure, and skilled talent. Restrictions affecting cross-border data use and cloud services, can make it more difficult for organizations to train, test, improve, and deploy AI systems at scale. Enabling policy environments that promote interoperable governance approaches and the practical enablers of diffusion (e.g. compute, data, skills development, collaborative experimentation models) can help support responsible AI adoption in practice. The private sector is already contributing to addressing these gaps through the implementation of governance and risk management practices at scale, as well as through partnerships that support capacity building, skills development, and deployment-focused collaboration that helps translate governance into practice. The Dialogue provides an opportunity to bring forward this practical experience through case studies and evidence that can help inform policy discussions and identify areas where further coordination is needed. A key opportunity is to use the Dialogue to highlight existing tools, partnerships, and programmes that can be scaled or adapted across contexts. More predictable and interoperable approaches could enable a wider range of businesses, including SMEs, to adopt, develop, or deploy AI systems, and help ensure that the benefits of AI can be more widely shared, including for developing and emerging economies where strengthening capacity for adoption and implementation – not only regulation – is a central priority.

The Dialogue's is in the unique position to serve as a connector and synthesiser across existing international AI governance efforts, rather than adding another parallel process. There is already a rich ecosystem of frameworks, principles, and initiatives, and one of the Dialogue's most valuable functions would be to build a shared understanding of how AI is deployed in practice and helping to address existing fragmentation and fill evidence gaps that currently limit effective policymaking. The Dialogue is well placed to strengthen the evidence base for governance by grounding discussions in real-world use cases, implementation challenges and lessons learned, and by translating technical insights into actionable policy understanding. Focusing on capacity to act, rather than solely on capacity to regulate, will be especially important for developing and emerging economies, where the priority is often to enable responsible AI adoption. AI governance debates continue to move faster than the available evidence on how systems are designed, developed, and used, particularly across different regions and sectors. In this context, the Dialogue should enable continuous information-sharing focusing on the practical enablers of AI diffusion, including public-private partnerships that support responsible uptake of AI. Rather than creating new governance structures or duplicative capacity building initiatives, the Dialogue should highlight, connect and scale existing programmes, partnerships, and tools that are already delivering impact.

The Dialogue should actively build on existing frameworks and processes rather than starting from scratch. These include, but are not limited to: UNESCO's 2021 Recommendation on the Ethics of AI, the OECD AI Principles, the OECD AI Policy Observatory, the Internet Governance Forum Policy Network on AI, the G7 Hiroshima AI Process, the G20 AI work, and the AI Safety Summits hosted so far in the UK, South Korea, France, India, and the upcoming 2027 edition in Switzerland, that can help shape the Dialogue's evidence base and agenda-setting. In terms of process, the Dialogue can learn from models that have successfully structured multistakeholder engagement over time, such as the Internet Governance Forum, the World Summit on the Information Society (WSIS) Forum and the São Paulo Multistakeholder Guidelines. Both have developed practical experience in structuring inclusive, ongoing exchange over time, including mechanisms (e.g. IGF Multistakeholder Advisory Group (MAG), intersessional workstreams, open consultations) that extend engagement beyond annual meetings and encourage continuous stakeholder input. Practically, the Dialogue could add value by developing shared tools, such as a publicly accessible repository of AI governance practices, searchable by sector and geography, that allow stakeholders to learn from each other's experience and build cumulatively on what works. To maximise impact, the Dialogue should focus on complementary, evidence based contributions that strengthen understanding of how AI is deployed in practice, rather than revisiting high level principles already discussed in other fora. Strong linkages with major AI and digital policy convenings, both within and outside the UN system, will be essential to ensure coherence, share insights, and allow external developments to inform the Dialogue's work over time.

Different stakeholders can bring distinct but complementary strengths to the Global Dialogue on AI Governance. Business contributes operational expertise and practical know-how from developing and deploying AI systems at scale; the technical community brings system level and engineering perspectives; civil society provides human rights and societal insights; academia contributes scientific rigour and forward-looking research; and governments contribute institutional perspectives on policy implementation and coordination. Harnessing these perspectives effectively is key to making the Dialogue impactful. Contributions should focus on real world use cases and operational experience that help policymakers understand what enables, and what constrains, responsible AI diffusion in practice. Businesses can provide concrete examples of how AI is deployed across sectors and regions, highlighting best practices and operational blockers. These insights are particularly valuable in informing how policy environments can be made more enabling, especially for SMEs and actors in developing and emerging economies. To facilitate this exchange, the Dialogue should prioritise practical, solution-oriented formats that move beyond prepared statements. This could include mixed panels, thematic roundtables, and sector-specific deep-dive sessions that bring together policymakers, industry, and other stakeholders to discuss one concrete challenge at a time and identify actionable takeaways. Grounding discussions in practical experience will help ensure that the Dialogue informs realistic, effective policy outcomes. Between annual sessions, light-touch and complementary intersessional mechanisms would help maintain continuity and ensure that input is not confined to the July or May Dialogues. Such arrangements would also provide a clear bridge between the first and second editions of the Dialogue.

The perspectives of businesses from developing and emerging economies, particularly small and medium-sized enterprises that lack the resources to engage with international processes, could be more consistently reflected in global AI governance discussions. Meaningful inclusion of these stakeholders requires that the Dialogue reflects their priorities and practical realities. For many actors in developing economies, the primary challenge is often not the absence of regulatory frameworks, but limited capacity to adopt, deploy, and benefit from AI in practice. This highlights the importance of complementing discussions on governance frameworks with greater attention to capacity to act, including access to data, compute resources, skills, infrastructure, and interoperable policy environments that support cross-border collaboration and scaling. Broader participation can also be supported through active outreach, logistical support (including funding for travel and participation costs where relevant), and formats that are genuinely accessible, in terms of language, timing, and the level of assumed familiarity with UN processes. Creating space for diverse stakeholder perspectives, including stakeholders from Least Developed Countries (LDCs) and underrepresented groups, can help ensure that discussions reflect a wider range of implementation experiences and development contexts.

Innovative engagement formats will be essential to foster meaningful and dynamic participation to ensure that the Dialogue produces practical, actionable outcomes. Issue-based working sessions, structured around a specific and clearly bounded question rather than broad themes, can help participants with diverse backgrounds engage on equal footing and focus discussions on problem-solving. Linking these sessions to short, publicly accessible synthesis notes would help build a shared evidence base over time and ensure continuity across Dialogue editions. Sector-based working sessions can further strengthen engagement by bridging technical, operational and policy perspectives. These formats allow businesses to explain how AI is deployed in real-world settings, while enabling policymakers and other stakeholders to better understand the operational constraints, enabling conditions, and trade-offs faced in practice. Such exchanges are particularly effective in generating transferable insights that can inform more coherent and interoperable governance approaches across sectors and regions. To encourage meaningful, practice-oriented contributions from industry, the Dialogue should prioritise interactive and solution-oriented formats over prepared interventions. Sessions should be designed to facilitate the sharing of best practices and the exchange of implementation experience, with the aim of identifying barriers to AI adoption and diffusion and highlighting what works in practice. Discussions should focus on the concrete conditions that enable responsible AI adoption, including access to compute, high-quality data, skills development, and collaborative experimentation models. Deployment-focused partnerships can serve as useful reference points for accelerating learning and responsible uptake. Complementing these formats, the Dialogue should also provide space for bilateral or small-group exchanges to support more targeted discussions between stakeholders. Existing stakeholder coalitions and multistakeholder platforms can be leveraged to help convene such interactions.

From a multistakeholder process perspective, the Internet Governance Forum (IGF) provides a strong example. The IGF's open, inclusive, and year-round engagement model, supported by its intersessional work and national and regional IGF initiatives, has demonstrated how global discussions can be informed by local and regional realities. In addition, the work of the Multistakeholder Advisory Group (MAG) in shaping the IGF programme and agenda helps ensure balanced representation of views and priorities across stakeholder groups and regions, while maintaining flexibility to address emerging issues. This approach offers a tested model for structuring inclusive agenda-setting and sustained engagement beyond a single annual meeting. In terms of policy tools and practices, the OECD AI Policy Observatory illustrates how a shared, publicly accessible repository of national AI policies, regulatory approaches and initiatives can support comparative learning and evidence-based policymaking. Importantly, it does so without requiring harmonisation, allowing jurisdictions to learn from one another while respecting different legal, institutional and developmental contexts. Together, these initiatives highlight the value of platforms that prioritise transparency, accessibility, and cumulative learning. The Global Dialogue could add value by acting as a connector across these efforts, helping to map existing frameworks, surface best practices, and identify gaps or points of friction, particularly in relation to the interoperability of AI governance approaches. This is increasingly important in light of growing regulatory fragmentation, which can constrain cross-border cooperation and slow AI diffusion.