Ilustre y Nacional Colegio de Abogados

A successful first Global Dialogue on AI Governance would be one that moves beyond general principles and contributes to greater clarity on how governance frameworks can be effectively implemented in practice. In particular, success would involve advancing shared understanding on how to ensure accountability in increasingly complex, multi-actor AI ecosystems, where responsibilities are often diffused across States, private sector actors, and other non-State entities. This includes identifying practical approaches to clarify roles, strengthen oversight mechanisms, and ensure that the development and deployment of AI systems remain consistent with international law and existing legal obligations. From this perspective, progress in defining transparent, accountable, and interoperable governance approaches would be essential, especially those that can be adapted across different jurisdictions while maintaining coherence. Equally important is ensuring that these frameworks are capable of preventing harm, addressing legal responsibility, and enabling effective remedies, particularly for individuals and communities most vulnerable to the risks associated with AI systems. Finally, the Dialogue would be successful if it helps bridge the gap between policy discussions and real-world implementation, by promoting capacity-building efforts that integrate legal, institutional, and risk-informed perspectives into AI governance.

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

My selection reflects the need to address artificial intelligence not only as a technological development, but as a source of evolving legal and governance challenges that require urgent and coordinated responses. The priority given to safe, secure and trustworthy AI responds to the increasing integration of AI systems into critical decision-making processes, where failures, misuse, or vulnerabilities may generate systemic risks with legal and societal implications. Ensuring security and reliability is therefore essential to prevent harm and maintain institutional trust. The focus on the protection and promotion of human rights reflects the centrality of individuals in AI governance. As AI systems influence access to services, information, and opportunities, it is critical to ensure that their design and deployment uphold fundamental rights, particularly for those in vulnerable situations. Finally, transparency, accountability, and human oversight are key to addressing the complexity of multi-actor AI ecosystems. Clarifying responsibilities across developers, deployers, and users, and ensuring meaningful human oversight, are necessary conditions to enable legal accountability, prevent misuse, and support effective remedies. Together, these priorities contribute to a governance approach that is both practical and grounded in the rule of law.

In my view, an important cross-cutting issue that is not sufficiently captured in the current thematic areas is the use of artificial intelligence by organized criminal groups, terrorist actors, and, in some contexts, by States to enable or amplify high-impact illicit activities. AI is increasingly being leveraged to enhance capabilities such as large-scale fraud, cyber-enabled extortion, identity theft, disinformation campaigns, and the automation of malicious operations. The use of generative AI to produce deepfakes, synthetic identities, or targeted manipulation content significantly lowers barriers to entry and increases the scale, speed, and sophistication of these activities. In parallel, the potential misuse of AI tools in intelligence, surveillance, or influence operations raises complex legal and governance questions, particularly when such uses may indirectly facilitate harm or undermine public trust. This evolving landscape highlights the need to better integrate security, criminal, and risk perspectives into AI governance discussions, beyond traditional regulatory or ethical frameworks. It also underscores the importance of clarifying responsibility across multi-actor environments, where developers, providers, users, and State actors may be interconnected in ways that complicate accountability. Addressing these dynamics requires strengthening cooperation, improving legal and institutional responses, and ensuring that governance frameworks are capable of anticipating and mitigating the misuse of AI in contexts of serious and organized harm, while protecting individuals and reinforcing the rule of law.

In Latin America, governance gaps in safe, secure and trustworthy AI, protection and promotion of human rights, and transparency, accountability, and human oversight are particularly visible because our region is primarily a consumer of technologies developed elsewhere. As a result, deployment often outpaces understanding, oversight, and adaptation to local legal and institutional realities. Across public institutions, private sector actors, and society at large, the rapid and widespread adoption of AI is not always accompanied by sufficient awareness of risks, responsibilities, and legal implications. The technology evolves at an exponential pace, while legal frameworks, regulatory capacity, and technical understanding develop more slowly. This gap creates uncertainty in areas such as liability, due process, data protection, and the effective exercise of rights. A key challenge is that many organizations integrate AI tools without fully understanding their limitations, biases, or potential impacts, which can lead to unintended harm or systemic vulnerabilities. In parallel, limited local capacity and fragmented governance frameworks can make it difficult to enforce accountability or ensure meaningful human oversight. These dynamics are compounded by asymmetries in the global digital ecosystem, where large technology providers shape standards and practices that may not fully reflect local contexts or needs. At the same time, this presents an opportunity to strengthen capacity-building, legal interpretation, and institutional coordination, enabling more informed adoption of AI and more effective alignment with international principles, while reinforcing the rule of law in rapidly evolving digital environments.

The AI Dialogue can play a critical role by shifting international cooperation from reactive discussions toward forward-looking and anticipatory governance. In a context where technological development advances at an exponential pace, dialogue focused only on present challenges often leads to solutions that arrive too late or are already outdated. For this reason, the Dialogue should promote the development of future-oriented scenarios that allow governments and stakeholders to better understand emerging risks, anticipate impacts, and plan more effectively in the present. At the same time, the Dialogue can serve as a space for frank and transparent exchanges on the role and interests of major technology companies, whose influence in shaping AI systems and standards is increasingly significant. Addressing these dynamics openly is essential to ensure that governance approaches are not only technically sound, but also balanced and aligned with broader public interests. In a global environment where multilateralism faces growing challenges, the Dialogue has the opportunity to reinforce cooperation by fostering trust, inclusiveness, and shared understanding across diverse actors. By connecting legal, technical, and policy perspectives, it can help build more coherent and adaptive frameworks that respond to rapid technological change while supporting accountability, human rights, and the rule of law.

The AI Dialogue could add significant value by first undertaking a systematic mapping of existing initiatives, partnerships, and governance mechanisms, rather than prioritizing alignment with any single framework. Given the diversity of ongoing efforts at regional, sectoral, and global levels, a structured approach to identifying and understanding these initiatives would help avoid fragmentation and duplication. Such a mapping could be guided by a set of common indicators, including scope of application, geographic reach, thematic focus, level of innovation, degree of implementation, and relevance to different stakeholders. This would allow the Dialogue to identify complementarities, gaps, and areas where coordination could be strengthened. By serving as a platform that brings together these diverse efforts under a more coherent analytical framework, the AI Dialogue could facilitate greater interoperability, mutual learning, and practical cooperation. Its added value would lie not in replacing existing initiatives, but in connecting them through a more integrated and forward-looking perspective, helping to translate dispersed knowledge into more consistent and actionable approaches to AI governance.

Different stakeholders can contribute to the AI Dialogue by bringing evidence-based knowledge, practical experience, and independent analysis into the discussion. This includes sharing research, technical insights, policy lessons, and case studies that reflect how AI systems are being developed, deployed, and governed across different contexts. Beyond this, stakeholders have a critical role in helping to anticipate emerging risks and opportunities, contributing to the development of forward-looking scenarios that allow the Dialogue to move beyond present challenges and better prepare for future developments. In this regard, stakeholders should not only inform the conversation, but also help to "imagine the unimaginable"—exploring how AI could reshape security, governance, and societal dynamics in ways that are not yet fully understood. From a structural perspective, the Dialogue would benefit from a more dynamic and inclusive format. While maintaining the necessary level of formality, it should move beyond traditional, rigid formats that often limit meaningful exchange. Discussions could be designed to interleave speakers from different sectors—governments, private sector, academia, and civil society—ensuring balanced participation and avoiding overconcentration of speaking time. In technological discussions, all stakeholders should contribute on a more equal footing. This would strengthen the quality of the exchange, enhance mutual learning, and ensure that diverse perspectives meaningfully shape outcomes.

Voices from vulnerable communities, particularly indigenous peoples, remain significantly underrepresented in global discussions on AI governance. This is not always due to a lack of invitation, but rather to structural barriers that limit meaningful participation. In many cases, these communities already interact with digital technologies—sharing data, using platforms, and increasingly becoming targets of fraud or manipulation—without fully understanding the implications or risks involved. For indigenous communities, these challenges are compounded by linguistic and cultural gaps. If engagement is already difficult at the level of national languages, the additional layer of technical language associated with AI further excludes them from informed participation. As a result, they are often affected by technologies they neither shape nor fully understand. To address this, inclusion efforts should go beyond formal participation and focus on accessible communication, localized capacity-building, and culturally appropriate engagement strategies. This includes translating key concepts, developing community-oriented education initiatives, and creating safe spaces for dialogue. Ensuring their meaningful participation is essential to building AI governance frameworks that are truly inclusive and responsive to real-world impacts.

Innovative engagement formats for the AI Dialogue could combine artificial intelligence tools with foresight methodologies to create more interactive, forward-looking discussions. One approach would be to use scenario-based simulations, where participants engage with plausible future situations involving AI risks, governance failures, or emerging opportunities. These scenarios could be supported by AI tools that generate dynamic variables, allowing participants to explore how different decisions may lead to different outcomes. Additionally, the Dialogue could incorporate real-time AI-assisted synthesis, where key ideas, points of convergence, and disagreements are mapped live during discussions. This would enhance transparency and help structure complex conversations more effectively. Another useful format would be multi-stakeholder foresight labs, where small, diverse groups work collaboratively to identify risks, test governance approaches, and propose actionable recommendations. These labs could be designed to integrate perspectives from governments, industry, academia, and civil society on an equal footing. Such formats would move the Dialogue beyond traditional interventions, fostering deeper engagement, practical insight, and more adaptive governance thinking.

In the Latin American context, it is difficult to identify policies or practices that stand out as comprehensive or particularly effective examples of AI governance. While there are ongoing initiatives and regulatory discussions, few have reached a level of maturity, implementation, or impact that could be considered a strong reference at the global level. This situation should be a matter of concern. It reflects existing gaps in capacity, coordination, and long-term strategic vision, and highlights the need to strengthen institutional frameworks, develop context-specific approaches, and accelerate efforts to translate principles into practical, enforceable governance mechanisms.