Panama´s Secretariat of Science, Technology and Innovation

Panama considers the first Global Dialogue a success if it produces three concrete outcomes. First, an agreed thematic agenda for 2027 that reflects the priorities of developing countries, not only those of large AI-producing states. Second, a clear framework for voluntary reporting on national AI governance approaches, enabling genuine knowledge-sharing without imposing prescriptive standards on countries with different capacities. Third, the launch of a structured technical assistance mechanism — co-designed with regional bodies such as ECLAC, CAF, and the IDB — that links governance commitments to implementable support for nations still building foundational AI infrastructure. The Dialogue should generate interoperable frameworks and capacity-building pipelines that allow countries like Panama to advance implementation. Finally, a successful Dialogue will generate multi-stakeholder buy-in by giving private sector, civil society, and scientific institutions a structured role in shaping governance outcomes.

• Safe, secure and trustworthy AI
• AI capacity-building
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches

Safe, secure and trustworthy AI: Panama's ENAI adopts a risk-based approach aligned with international best practices. Trustworthiness is a prerequisite for AI adoption in public services where Panama is actively deploying AI tools. Without verifiable safety standards, public trust erodes and adoption stalls. AI capacity-building: Panama does not yet have large-scale AI research infrastructure. The country depends on international partnerships and funding mechanisms to develop talent, compute access, and data governance frameworks. The Dialogue must make capacity-building a first-tier issue. Interoperability of governance approaches: Panama engages simultaneously with the OAS COMCYT, UNESCO, the OECD AI Policy Observatory, and bilateral partners. Divergent governance frameworks create compliance complexity for small administrations. The Dialogue should map existing frameworks and identify convergence zones that reduce duplication without mandating harmonization. Social, economic, ethical, cultural, and linguistic implications: Panama is a multicultural, multilingual society with indigenous populations and significant linguistic diversity. AI systems trained predominantly on high-resource languages create equity deficits. Panama advocates for governance approaches that explicitly address linguistic inclusion and the distribution of AI economic benefits.

1. AI and strategic infrastructure security: Panama's Canal and logistics corridor are critical global chokepoints. The integration of AI into port operations, maritime traffic management, and logistics systems introduces new cybersecurity and operational risk vectors. Existing AI governance frameworks do not adequately address the intersection of AI, critical infrastructure, and geopolitical exposure. 2. Environmental costs of AI: Training large models and operating AI data centers carry significant energy and water footprints. For countries with ambitious environmental commitments, the governance of AI's physical resource demand is a sovereignty and sustainability issue, not only a technical one. 3. AI and scientific diplomacy: The governance of AI-enabled scientific collaboration - including data sovereignty in joint research, IP rights in AI-assisted discovery, and access to frontier models for public-good research - is absent from current multilateral agendas. Panama, as an active participant in international scientific cooperation, has a direct interest in this gap being closed.

Panama operates in a governance environment with significant asymmetry. Major AI-producing countries set de facto standards through their own regulatory frameworks (EU AI Act, U.S. Executive Orders, China's AI regulations), while developing countries must absorb these standards without having shaped them. This creates a compliance burden for small administrations and risks locking Panama into governance models designed for different legal, institutional, and economic contexts. The most significant challenge is the gap between AI governance declarations and implementable policy tools. Panama is moving towards enshrining experimentation and the use of innovative regulation approaches (such as sandboxes) to cope with a dynamic international regulatory environment and domestic institutional bandwidth constraints.

For Panama, the Dialogue's most valuable function is as a norm-translation mechanism. It should convert broad principles — such as those in the UNESCO Recommendation on the Ethics of AI or the Bletchley Declaration — into implementable governance tools that countries with limited administrative capacity can adopt. This requires the Dialogue to produce practical guidance, model regulations, and assessment frameworks. The Dialogue should also serve as an early warning system for governance gaps. Rapid developments in frontier AI — agentic systems, multimodal models, AI in biosecurity — outpace current regulatory frameworks. A standing multilateral dialogue with a defined review cycle can identify emerging risks and coordinate responses before fragmented national regulation creates inconsistencies that harm global interoperability. Finally, the Dialogue can function as a trust-building mechanism between AI-leading and AI-receiving countries, creating structured channels for dialogue on compute access, data governance, and technology transfer that do not currently exist in any multilateral forum.

The Dialogue should actively connect with the following existing frameworks to avoid duplication and leverage established trust networks: UNESCO Recommendation on the Ethics of AI (2021): The most broadly endorsed multilateral AI ethics instrument. The Dialogue should use it as a baseline for governance principles and build upon its implementation toolkit. OECD AI Principles and AI Policy Observatory: The Observatory provides the most comprehensive comparative database of national AI policies. The Dialogue should integrate its data to ground governance discussions in evidence. ITU AI for Good Summit: The co-location of the first Dialogue with the AI for Good Summit is strategically sound. The Dialogue should establish a standing coordination mechanism with ITU to address technical standardization alongside governance. OAS COMCYT and ECLAC: For Latin America, these are the primary regional bodies with mandates in AI governance and digital economy. The Dialogue should designate them as formal regional consultation channels.

Panama recommends a multi-track structure for the Dialogue that distinguishes between negotiation-level engagement (member states) and advisory-level input (private sector, civil society, scientific community, technical bodies). For member states: Formal plenary sessions with structured intervention time distributed equitably across regional groups. Developed country delegations should not dominate floor time. Written submissions, like this one, should be synthesized and made publicly available as background documentation. For private sector: Structured engagement sessions where industry representatives respond to specific governance questions identified by member states. For civil society and scientific institutions: An independent advisory track with formal reporting rights into plenary sessions. Scientific should have a defined role in providing technical assessments of governance proposals.

The most systematically underrepresented voices in global AI governance are those of the Global South particularly small island developing states, least developed countries, and middle-income countries in Latin America, Africa, and Southeast Asia. These countries are among the most affected by AI's socioeconomic disruptions but have the least capacity to shape the governance frameworks that will govern its deployment. Specific underrepresented communities include: indigenous populations, whose languages, knowledge systems, and cultural heritage are largely absent from AI training datasets and from governance discussions; women and gender-diverse communities in technology policy roles in developing country contexts; and small and medium enterprises in the Global South, who are AI adopters but not AI developers, and face significant compliance burdens from divergent governance frameworks. Concrete inclusion mechanisms Panama recommends: (1) A fellowships program providing funding for civil society representatives from underrepresented regions to attend and prepare for Dialogue sessions. (2) Advance submission of country position papers with translation support into all six UN languages. (3) Rotating regional co-facilitation, ensuring that co-chairs in future Dialogue cycles include representatives from Africa, Asia-Pacific, and Latin America. (4) Remote participation infrastructure designed for low-bandwidth environments, not only for convenience in high-connectivity contexts.

Country case study presentations: Short, structured presentations by member states on specific governance implementations, successes, failures, and lessons learned. These are more actionable than general position statements and create genuine peer learning. Governance gap labs: Working groups formed around specific implementation challenges — e.g., how to conduct AI conformity assessments with limited regulatory capacity — that produce draft guidance documents for member state review. Persistent digital platform: Between physical sessions, a structured online deliberation platform where member states and accredited stakeholders can advance negotiations on specific text. This reduces the burden on physical negotiating sessions and allows smaller delegations to engage at their own pace.

Panama is developing an AI regulatory framework in collaboration with the ITU that is grounded in experimentation and adaptive regulatory approaches. The framework is designed to account for institutional capacity constraints while remaining responsive to new information and shifts in the technological landscape. This approach is currently unique in Latin America; its effectiveness remains under evaluation.