Asociación Chilena de Inteligencia Artificial para el Desarrollo Sostenible

For the first Global Dialogue on AI Governance to be considered a success, it must deliver outcomes that go beyond declarations of intent and translate into concrete commitments with measurable impact, particularly for regions that remain underrepresented in global AI governance discussions. From the perspective of ACHIADS, the Chilean Association for Artificial Intelligence for Sustainable Development, and based on our work promoting responsible and ethical AI adoption across Latin America, three outcomes would define success. First, the consolidation of a genuinely inclusive participation framework. The Dialogue must demonstrate that civil society organizations from the Global South, such as ACHIADS, are not peripheral voices but active co-designers of global governance frameworks. Regional associations with documented expertise in responsible AI, sustainability, and capacity building must have structured channels to contribute substantively to the Dialogue's outcomes. Second, the adoption of interoperability guidelines that are compatible with the realities of developing economies. ACHIADS has observed firsthand how national AI strategies in Latin America risk fragmentation and misalignment with international standards. The Dialogue should produce actionable guidance enabling coherence between regional frameworks and emerging global norms, without imposing models designed exclusively for high-income contexts. Third, a binding commitment to AI capacity building as a core governance priority. Asymmetric governance is ineffective governance. Organizations like ACHIADS invest significantly in building institutional, technical, and ethical AI capacity at the national level, but this work requires multilateral support. A successful Dialogue must produce a concrete roadmap for closing capacity gaps, including access to tools, training, and knowledge transfer mechanisms. Success will ultimately be measured by whether the Dialogue generates the trust and shared purpose needed for every country and every civil society organization, regardless of development level, to govern AI in alignment with their own values and the Sustainable Development Goals.

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI

ACHIADS selected these two thematic areas because they are deeply interconnected and reflect the most urgent governance challenges observed through our work in Chile and Latin America. Regarding safe, secure and trustworthy AI, our experience accompanying organizations in their AI adoption processes confirms that the absence of clear safety and trustworthiness standards is the primary barrier to responsible implementation, particularly in public institutions, small and medium enterprises, and civil society organizations. Without a shared international foundation for what constitutes safe and trustworthy AI, national efforts remain fragmented and vulnerable. ACHIADS actively promotes governance frameworks grounded in these principles, and we see the Dialogue as a critical opportunity to establish universally applicable baselines that are also adaptable to diverse regional contexts. Regarding the social, economic, ethical, cultural, linguistic and technical implications of AI, this area directly reflects ACHIADS's institutional mission. Latin America presents a unique combination of factors: significant socioeconomic inequality, rich cultural and linguistic diversity, and rapidly expanding AI adoption with limited regulatory infrastructure. The implications of AI in this context are not abstract. They affect labor markets, access to public services, democratic processes, and the preservation of cultural identity. Any governance framework that fails to account for this diversity will be inequitable by design. Together, these two priorities represent a coherent position: AI governance must simultaneously ensure that systems are safe and trustworthy at a technical level, and that their broader human implications are understood, monitored, and addressed with full recognition of regional realities. ACHIADS is committed to contributing evidence-based perspectives from Latin America to both dimensions of this agenda within the Global Dialogue.

From ACHIADS's perspective, there are at least three significant cross-cutting issues that the thematic areas identified in Resolution 79/325 do not fully capture and that deserve explicit attention within the Global Dialogue. First, AI sovereignty and strategic autonomy for developing nations. The current global AI landscape is characterized by a high concentration of foundational infrastructure, large language models, and governance standard-setting capacity in a small number of countries and corporations. This asymmetry creates structural dependency that undermines the ability of developing nations to govern AI in alignment with their own values, needs, and development priorities. Sovereignty in AI governance must become an explicit agenda item, not a secondary consideration. Second, the environmental and climate implications of AI. The rapid scaling of AI systems carries significant energy consumption and carbon footprint consequences that disproportionately affect developing regions already facing climate vulnerability. A governance framework aligned with the Sustainable Development Goals cannot treat environmental sustainability as peripheral to AI policy. This dimension requires dedicated attention, particularly regarding data center infrastructure, resource extraction for hardware, and the carbon costs of large-scale model training. Third, the governance of AI in the context of regional and linguistic diversity. While the listed themes reference cultural and linguistic implications, they do not adequately address the structural underrepresentation of non-English languages and non-Western epistemologies in AI systems. For Latin America, this translates into AI tools that frequently fail to reflect regional realities, local knowledge systems, and the linguistic richness of Spanish, Portuguese, and indigenous languages. Governance frameworks must actively address this gap, not as a cultural courtesy, but as a matter of epistemic equity. These three issues are not marginal. They are foundational to ensuring that the Global Dialogue produces governance outcomes that are genuinely universal and equitable.

Chile and Latin America are at a critical inflection point in AI governance. Through ACHIADS's work across public institutions, private sector organizations, and civil society, we observe both significant challenges and concrete opportunities directly linked to the governance gaps in our two priority thematic areas. Regarding safe, secure and trustworthy AI, the primary challenge is the absence of binding national frameworks in most Latin American countries. Chile has made meaningful progress with its National AI Policy and ongoing regulatory discussions, but implementation remains uneven and enforcement capacity is limited. Organizations adopting AI systems frequently do so without adequate risk assessment tools, transparency standards, or accountability mechanisms. This gap is particularly acute in high-stakes sectors such as healthcare, education, and public administration, where untrustworthy AI systems can cause disproportionate harm to vulnerable populations. Regarding the social, economic, ethical, cultural, linguistic and technical implications of AI, the challenges are structural. Labor market disruption from AI automation is advancing faster than policy responses, particularly affecting middle-skill workers in sectors with limited retraining infrastructure. Additionally, AI systems deployed in the region frequently embed biases reflecting realities far removed from Latin American social and cultural contexts, generating discriminatory outcomes that erode public trust. However, significant opportunities exist. Latin America has a growing ecosystem of civil society organizations, research institutions, and government bodies actively developing regional AI governance capacity. Chile's participation in international AI governance forums, combined with the work of organizations such as ACHIADS in promoting responsible adoption frameworks grounded in sustainability and ethics, positions the region to contribute meaningfully to global governance discussions rather than merely receiving externally designed standards. The Global Dialogue represents a concrete opportunity to transform these regional efforts into inputs that shape universally applicable and contextually sensitive governance frameworks.

The Global Dialogue on AI Governance has the potential to play a transformative role in international cooperation, provided it operates as a genuinely multilateral mechanism rather than a forum where governance standards designed in high-income contexts are ratified with broad participation but limited co-authorship. From ACHIADS's perspective, the Dialogue can advance international cooperation in three concrete ways. First, by serving as a bridge between existing governance initiatives and underrepresented regions. Multiple AI governance frameworks currently coexist, including the OECD AI Principles, the Hiroshima Process, the EU AI Act, and various national strategies. However, Latin America, Africa, and Southeast Asia remain largely rule-takers in this landscape. The Dialogue can create structured mechanisms for regional organizations and civil society bodies to contribute their governance experience and contextual knowledge, ensuring that international standards reflect diverse realities rather than consolidating existing asymmetries. Second, by establishing a shared evidence base through its relationship with the Independent International Scientific Panel on AI. Effective international cooperation requires common ground on facts. The Panel's independent scientific assessments can provide the Dialogue with a neutral foundation for discussions that are frequently distorted by geopolitical interests or corporate influence. ACHIADS strongly supports ensuring that Panel findings are accessible, translated, and actively disseminated across all regions, including in Spanish and Portuguese. Third, by generating binding or semi-binding cooperation instruments on capacity building. Declarations of solidarity are insufficient. The Dialogue must produce concrete cooperation mechanisms, including technology transfer agreements, joint training programs, and shared governance toolkits, that allow organizations like ACHIADS to strengthen national and regional AI governance capacity with multilateral support. International cooperation on AI governance will only be meaningful if it is built on equity, shared evidence, and instruments that reduce rather than reproduce existing global inequalities.

Several existing initiatives provide a valuable foundation upon which the Global Dialogue should build, while also identifying clear gaps where the Dialogue can generate distinct added value. At the global level, the OECD AI Principles and the Hiroshima AI Process have established important normative references for responsible AI governance. The UNESCO Recommendation on the Ethics of AI represents the broadest multilateral consensus achieved to date on ethical principles. The Global Dialogue should build upon these frameworks while ensuring they are translated into actionable governance instruments accessible to countries and organizations beyond the OECD membership. At the regional level in Latin America, initiatives such as the Inter-American Development Bank's AI governance programs, ECLAC's work on digital transformation and the digital economy, and the emerging network of national AI observatories represent concrete institutional infrastructure. ACHIADS actively engages with this ecosystem and can attest to its growing capacity and legitimacy. The Dialogue should formally recognize and connect with these regional mechanisms rather than creating parallel structures that fragment existing efforts. At the civil society level, organizations such as ACHIADS contribute evidence-based, practitioner-grounded perspectives on responsible AI adoption that complement the policy and technical focus of intergovernmental processes. Structured civil society engagement mechanisms within the Dialogue would allow this knowledge to inform governance outcomes systematically. The distinct added value the Global Dialogue can bring is threefold: first, universal legitimacy as a United Nations platform that no regional or sectoral initiative can replicate; second, the ability to connect the Independent International Scientific Panel's evidence directly to policy deliberation; and third, the capacity to generate governance coordination across initiatives that currently operate in parallel without sufficient coherence. The Dialogue's greatest contribution will be transforming a fragmented global governance landscape into a coherent, inclusive, and evidence-driven architecture.

Ensuring meaningful and diverse stakeholder participation is not a procedural matter; it is a fundamental condition for the legitimacy and effectiveness of the Global Dialogue. Based on ACHIADS's experience facilitating multi-stakeholder AI governance processes in Chile and Latin America, we offer the following recommendations. Regarding stakeholder contributions, governments must provide the formal decision-making backbone of the Dialogue, but their contributions will be stronger when informed by structured inputs from civil society, academia, and the private sector. Civil society organizations, particularly those operating at national and regional levels in the Global South, bring practitioner knowledge and community perspectives that are absent from purely intergovernmental processes. Academic and research institutions contribute independent technical analysis. The private sector brings implementation experience, provided its participation is governed by clear conflict-of-interest safeguards. Regarding format and structure, the Dialogue should adopt a tiered participation model. Plenary sessions should establish shared priorities and validate outcomes, while thematic working groups should allow deeper technical and policy deliberation with broader stakeholder participation. Regional preparatory forums, held prior to the annual Dialogue sessions, would ensure that perspectives from Latin America, Africa, and other underrepresented regions are consolidated and formally submitted as regional inputs rather than fragmented individual contributions. ACHIADS is prepared to actively participate in and help coordinate such a regional preparatory process for Latin America. Additionally, the Dialogue must address structural participation barriers. These include language accessibility, with documentation and deliberation available in all official UN languages and key regional languages; financial support mechanisms for civil society organizations from developing countries to attend and prepare contributions; and digital participation options that allow engagement beyond those who can travel to Geneva or New York. Inclusive participation is not achieved by opening the door. It requires actively removing the barriers that prevent meaningful engagement from the majority of the world.

Global AI governance discussions systematically underrepresent a significant majority of the world's population. From ACHIADS's perspective, the following voices and communities require urgent and structural inclusion. First, civil society organizations from Latin America, Africa, and Southeast Asia. These regions collectively represent the majority of the global population and are among those most exposed to the social and economic implications of AI adoption, yet their participation in standard-setting processes remains marginal. Inclusion requires not only open registration processes but dedicated representation quotas, regional coordination mechanisms, and financial support for participation. Second, indigenous communities and holders of traditional knowledge. AI systems increasingly interact with cultural heritage, linguistic diversity, and community data in ways that affect indigenous peoples directly. These communities are rarely consulted in AI governance processes, despite having legitimate claims over the data, languages, and knowledge systems that AI models appropriate. The Dialogue must establish formal consultation mechanisms that respect the principle of free, prior, and informed consent. Third, workers and labor organizations. The economic implications of AI automation fall disproportionately on working populations, yet labor unions and worker representatives are almost entirely absent from global AI governance forums. Their practical knowledge of how AI is transforming workplaces is indispensable for designing governance frameworks that protect labor rights and promote just transitions. Fourth, women and gender-diverse communities, particularly from developing regions. Gender bias in AI systems is well documented, yet women's organizations and gender equity advocates remain underrepresented in technical and policy governance spaces. To include these voices effectively, the Dialogue must move beyond tokenistic participation. This requires permanent structural mechanisms, including dedicated advisory bodies, regional liaison networks, and accountability processes that measure whether underrepresented perspectives have genuinely influenced governance outcomes, not merely been heard. ACHIADS commits to actively supporting these inclusion efforts within the Latin American region.

Effective engagement in the Global Dialogue requires moving beyond traditional conference formats that privilege prepared statements over genuine deliberation. Based on ACHIADS's experience designing and facilitating multi-stakeholder AI engagement processes in Latin America, we recommend the following innovative formats. First, structured deliberative sessions replacing traditional panel presentations. Rather than sequential expert presentations followed by brief question periods, the Dialogue should adopt facilitated deliberation formats where participants from different stakeholder groups and regions engage directly with specific governance challenges, producing concrete outputs such as priority matrices, consensus statements, or identified points of divergence that formally inform the Dialogue's conclusions. Second, regional input integration mechanisms. Prior to each annual session, regional preparatory forums should consolidate perspectives into formal regional position documents. These documents should be presented at the opening of the Dialogue as substantive inputs, not background materials, ensuring that the deliberation in Geneva or New York begins from a foundation of diverse, pre-organized regional knowledge. Third, open policy laboratories. Dedicated working sessions where participants collaboratively draft governance language, test regulatory approaches against real case studies, or evaluate the applicability of existing frameworks to diverse regional contexts would generate practical governance outputs with broad legitimacy. ACHIADS has applied similar methodologies at the national level in Chile with measurable results. Fourth, asynchronous digital participation platforms. A structured online engagement mechanism, active between annual sessions, would allow organizations that cannot attend in person to contribute analysis, respond to consultation documents, and participate in thematic working groups on a continuous basis. This is particularly relevant for civil society organizations in developing regions facing travel and resource constraints. Finally, youth and emerging leaders tracks with substantive, not ceremonial, participation would ensure intergenerational perspectives inform governance decisions whose consequences will be felt most acutely by future generations. Meaningful engagement is a design choice. The Dialogue must invest in formats that reflect that commitment.

Drawing on ACHIADS's direct experience accompanying AI governance processes in Chile and observing regional developments across Latin America, we highlight the following policies, practices, and approaches as relevant references for the Global Dialogue. At the national policy level, Chile's National Artificial Intelligence Policy represents a meaningful example of a developing country constructing a governance framework that balances innovation promotion with ethical principles and public interest safeguards. Its emphasis on human-centered AI, transparency, and alignment with the Sustainable Development Goals offers a replicable model for other Latin American nations building governance capacity from limited institutional starting points. At the organizational practice level, ACHIADS has developed and applied the ARES-AI Framework, a five-phase methodology for responsible AI adoption in organizations that integrates ethical assessment, stakeholder participation, sustainability criteria, and continuous monitoring. This framework has been implemented across diverse sectors in Chile, demonstrating that responsible AI governance is operationally viable beyond large corporations and government agencies, including in civil society organizations, cooperatives, and small and medium enterprises. At the regional level, the growing network of Latin American AI observatories and the Inter-American Development Bank's responsible AI initiatives provide infrastructure for evidence gathering, policy benchmarking, and cross-country learning that the Global Dialogue should formally recognize and connect with. At the multilateral level, the UNESCO Recommendation on the Ethics of AI stands as the most broadly inclusive normative reference available and has proven valuable as a foundation for national policy development in contexts with limited prior governance infrastructure. The common thread across these examples is that effective AI governance combines clear ethical principles with practical implementation tools, sustained stakeholder engagement, and mechanisms for continuous adaptation as technology evolves. The Global Dialogue should prioritize translating normative consensus into accessible, implementable governance instruments that organizations at all levels can adopt.