Fundapi

A successful first Global Dialogue on AI Governance would produce a clear and focused Co-Chairs' Summary that identifies a limited set of actionable priorities, helping move from broad principles to practical implementation. This could include shared reference elements for AI governance in areas such as data governance, transparency in public sector AI, and strengthening institutional capacity. It would also be important that the Dialogue meaningfully reflects the perspectives of countries at different levels of technological maturity, recognizing structural asymmetries in access to data, infrastructure, and AI development capabilities, and highlighting the need for more equitable participation in AI ecosystems. At the same time, the Dialogue should contribute to greater coherence across existing initiatives and frameworks, helping reduce fragmentation and identify opportunities for alignment. Its success would also depend on establishing continuity mechanisms beyond the event itself, such as thematic follow-up spaces that allow stakeholders to exchange experiences, tools, and lessons learned over time. Ultimately, the Dialogue would be successful if it supports a transition toward more coordinated, inclusive, and implementable approaches to AI governance, particularly for stakeholders operating in contexts with more limited resources and capacities.

• AI capacity-building
• Transparency, accountability, and human oversight
• Protection and promotion of human rights
• Open-source software, open data and open AI models

These priorities reflect a focus on ensuring that AI governance is both inclusive and implementable, particularly in contexts with more limited resources and institutional capacities, and are closely aligned with the work we carry out at Fundapi. AI capacity-building is essential to enable countries to meaningfully participate in the development and governance of AI systems, rather than remaining primarily as users of externally developed technologies. At Fundapi, we work on strengthening institutional and technical capacities in areas such as open data and digital transformation, which are foundational for effective AI governance. Transparency, accountability and human oversight are critical, especially in the use of AI within public decision-making. These principles directly connect with our work promoting open government, data transparency and citizen oversight mechanisms, ensuring that digital technologies remain subject to democratic control. Open-source software, open data and open AI models are key enablers for more equitable access to AI capabilities. Our experience supporting the development of open data ecosystems and public data use cases highlights how openness can reduce barriers, foster innovation and support more inclusive participation. Finally, the protection and promotion of human rights provides an essential overarching framework to guide AI governance. Our work emphasizes people-centered approaches to digital transformation, ensuring that technological adoption is aligned with rights, inclusion and public value. Together, these priorities reflect a practical approach that integrates capacity, openness, accountability and rights, and builds on our ongoing efforts to support more transparent, inclusive and effective governance processes.

Yes, while the listed themes are comprehensive, there are a few cross-cutting issues that could be more explicitly recognized. One key area is data governance as a foundational layer for AI. While aspects of data are implicitly covered, there is value in more clearly highlighting how data quality, interoperability, access, and stewardship underpin all AI systems. Strengthening data governance frameworks is particularly critical for countries seeking to build reliable, accountable and context-relevant AI applications. Another important issue is the concentration of AI development and infrastructure in a limited number of countries and actors. This structural imbalance affects not only access, but also the ability of many countries to shape AI systems according to their own social, cultural and policy contexts. Addressing this challenge requires not only capacity-building, but also more deliberate efforts to promote open ecosystems and more equitable participation in AI development. A third cross-cutting issue is the governance of AI in the public sector, particularly in relation to automated decision-making. This includes the need for standards, oversight mechanisms and safeguards to ensure transparency, accountability and public trust when AI is used in areas that directly affect citizens. Finally, there is an opportunity to more explicitly consider mechanisms for citizen participation and democratic oversight in AI governance. Beyond institutional frameworks, ensuring that individuals and communities can understand, question and influence how AI is used is essential for building legitimate and inclusive governance models. These issues cut across the existing themes and could help strengthen the Dialogue's focus on practical, equitable and context-sensitive implementation.

In Ecuador and across the region, governance gaps in the selected thematic areas are already shaping how AI is adopted and used, particularly in the public sector and in data-driven decision-making. A key challenge is limited institutional and technical capacity to design, procure and oversee AI systems. This often results in reliance on externally developed technologies, with limited ability to assess risks, ensure transparency or adapt systems to local contexts. At the same time, data governance frameworks are still evolving, which affects data quality, interoperability and the responsible use of information as a foundation for AI. Another significant gap relates to transparency, accountability and human oversight. While there is growing interest in using data and emerging technologies in government, there are still few established mechanisms to ensure explainability, auditability and citizen oversight, particularly when automated systems influence public services or policy decisions. Structural asymmetries in access to infrastructure, data and AI development capabilities also remain a major challenge. This reinforces a position where countries in the region are primarily consumers of AI, rather than active contributors, limiting opportunities for local innovation and context-sensitive solutions. At the same time, there are important opportunities. There is increasing momentum around open data, digital transformation and public sector innovation, which can serve as a foundation for more responsible AI governance. Civil society and multi-stakeholder initiatives are also playing a growing role in promoting transparency, participation and capacity-building. These dynamics highlight both the urgency and the opportunity to strengthen governance frameworks that are practical, inclusive and adapted to national realities.

The AI Dialogue can play a key role as a neutral, inclusive platform to bridge perspectives across countries, sectors and levels of development, helping to advance more coordinated and practical approaches to AI governance. First, it can support greater coherence in a fragmented global landscape by connecting existing initiatives, frameworks and regional efforts. By identifying areas of convergence and complementarities, the Dialogue can help avoid duplication and promote more aligned approaches, while still allowing flexibility for different national contexts. Second, it can facilitate the exchange of practical experiences and lessons learned, particularly by elevating voices from countries with different levels of technological maturity. This is essential to ensure that international cooperation is informed not only by advanced economies, but also by implementation realities in developing countries. Third, the Dialogue can help advance shared reference elements for AI governance that are adaptable and non-prescriptive. These could include approaches to data governance, transparency in public sector AI, and capacity-building pathways, providing a common foundation for cooperation without imposing uniform models. Finally, the Dialogue can play an important role in sustaining momentum by encouraging ongoing collaboration beyond the event itself. Establishing light follow-up mechanisms or thematic tracks would allow stakeholders to continue sharing tools, experiences and practices, gradually translating dialogue into implementation. Overall, the Dialogue can serve as a catalyst to move from fragmented discussions toward more inclusive, coordinated and action-oriented international cooperation on AI governance.

The AI Dialogue should build on and connect with existing global and regional initiatives that are already advancing principles, standards and implementation pathways for AI governance. These include the UNESCO Recommendation on the Ethics of Artificial Intelligence, which provides a comprehensive normative framework; the work of the OECD on AI principles and policy guidance; and regional efforts such as those led by the Economic Commission for Latin America and the Caribbean in promoting digital development and data governance in Latin America. It can also connect with multi-stakeholder initiatives focused on digital public infrastructure and open data ecosystems, as well as emerging national and regional AI strategies. In addition, partnerships that support capacity-building and technical cooperation, including those led by international financial institutions and development agencies, are critical to ensure that countries can translate governance principles into practice. Civil society networks and open government initiatives also play an important role in advancing transparency, accountability and citizen participation, which are central to AI governance. The added value of the AI Dialogue lies in its ability to act as a convening and bridging platform within the United Nations system, bringing together these diverse efforts into a more coherent and inclusive space. Rather than duplicating existing frameworks, it can help identify complementarities, surface practical lessons across regions, and elevate perspectives from countries with different levels of technological maturity. It can also contribute by promoting adaptable reference approaches and by fostering continuity through light coordination mechanisms that support ongoing collaboration and knowledge exchange.

Different stakeholders can contribute to the AI Dialogue by bringing complementary perspectives and practical experience. Governments can share policy approaches and regulatory priorities; the private sector can provide insights on technological development and implementation; academia and the technical community can contribute evidence and research; and civil society can bring perspectives on rights, accountability and real-world impacts. To make these contributions meaningful, the Dialogue should ensure balanced participation not only in speaking opportunities, but also in agenda-setting and follow-up processes, with particular attention to stakeholders from developing countries. In terms of format and structure, the Dialogue would benefit from strengthening interactive and practice-oriented components alongside high-level segments. Thematic discussions could be designed to prioritize concrete use cases and implementation experiences, allowing stakeholders to share lessons learned, challenges and solutions. This could be supported by short, focused interventions followed by moderated exchanges, rather than longer prepared statements. It would also be useful to ensure diversity within each thematic session, including representation from different regions, sectors and levels of technological maturity. Structured opportunities for written inputs before and after the Dialogue could help broaden participation beyond those able to intervene during the sessions. Finally, introducing light follow-up mechanisms linked to the thematic areas discussed—such as informal working groups or communities of practice—would allow stakeholders to continue contributing over time. This would help transform the Dialogue from a one-off event into an ongoing, collaborative process that supports practical progress in AI governance.

Several voices remain underrepresented in global discussions on AI governance, particularly those from developing countries, including Latin America, Africa and parts of Asia, where perspectives are often shaped by different institutional realities, resource constraints and social priorities. Within these contexts, civil society organizations, grassroots communities and public interest actors are not always adequately included, despite their critical role in highlighting real-world impacts and accountability concerns. Local governments and public sector practitioners are also frequently underrepresented, even though they are increasingly responsible for adopting and managing data-driven and AI-enabled systems. Their practical experience is essential to understanding implementation challenges, especially in areas such as service delivery, procurement and oversight. In addition, communities directly affected by AI systems—such as marginalized populations, indigenous groups, and those with limited digital access—are rarely present in these discussions, despite being among those most impacted by potential risks. To address these gaps, participation mechanisms should go beyond open calls and actively support inclusion through targeted outreach, regional consultations and partnerships with local organizations. Providing interpretation, flexible participation formats and financial support for participation can also help reduce barriers. Structurally, it would be beneficial to reserve space within discussions for underrepresented groups and ensure they are included not only as participants, but also as contributors to agenda-setting and follow-up processes. Strengthening connections with existing regional and local initiatives can also help bring more diverse perspectives into the Dialogue. Ensuring that these voices are meaningfully integrated is essential for building AI governance approaches that are inclusive, context-sensitive and responsive to real-world needs.

One effective approach would be use-case driven sessions, where stakeholders present short, structured examples of how AI is being implemented in specific contexts, followed by moderated discussion on lessons learned, challenges and implications for governance. This would ground the conversation in real-world experience and make discussions more actionable. Another useful format could be problem-solving labs or breakout groups, where participants work around a concrete policy or governance challenge, such as transparency in public sector AI or data governance frameworks. These sessions could bring together diverse stakeholders to collectively identify practical approaches or recommendations. Interactive polling and real-time feedback tools could also be used during plenary sessions to capture a broader range of perspectives and identify areas of convergence or divergence across participants. This would allow for more dynamic engagement beyond formal interventions. Additionally, multi-stakeholder roundtables with balanced representation could facilitate more inclusive dialogue, ensuring that voices from different regions, sectors and levels of technological maturity are actively heard. Short interventions followed by facilitated exchange can help maintain focus while encouraging interaction. Finally, incorporating pre- and post-dialogue engagement mechanisms, such as written inputs, online consultations or thematic discussion spaces, would extend participation beyond the event itself and allow stakeholders to contribute more substantively over time. Together, these formats can help make the Dialogue more participatory, grounded in practice and oriented toward generating useful and actionable outcomes.

From our experience at Fundapi, several practical approaches illustrate how AI governance can be advanced in ways that are accessible, transparent and aligned with public value. One example is the development of ComprasGPT, a tool designed to facilitate access to public procurement information. By simplifying complex datasets and enabling more intuitive interaction with procurement processes, this approach helps democratize participation, allowing a broader range of stakeholders to understand and engage with how public resources are allocated. This contributes to transparency and accountability in data-driven decision-making. In parallel, we have conducted research on algorithmic transparency in Ecuador, aimed at identifying gaps and raising awareness around the use of automated systems in public and private decision-making. This work has helped position the importance of transparency, explainability and human oversight, particularly in relation to the protection of human rights and the need for clear governance frameworks. Another key approach has been the development of accessible content and capacity-building initiatives, delivered through both online and offline formats. These efforts focus on translating complex topics such as data governance and AI into clear, practical knowledge for public officials, civil society and communities. This helps strengthen local capacities and supports more informed participation in governance processes. Together, these experiences highlight the importance of combining openness, practical tools, research and capacity-building to advance AI governance in a way that is inclusive, context-sensitive and oriented toward real-world impact.