Caribbean Telecommunications Union

A successful first Global Dialogue on AI Governance should establish a globally inclusive, action-oriented framework that meaningfully integrates the perspectives of developing regions, particularly Small Island Developing States (SIDS) such as those in the Caribbean; it can be defined by: 1. Recognizing structural asymmetries in AI development: the Dialogue must acknowledge disparities in infrastructure, financing, data access, and human capacity that limit meaningful participation by regions like the Caribbean, which risk remaining "standards-takers" rather than contributors to global AI governance 2. Committing to interoperable, regionally adaptable governance frameworks: global principles should support regional harmonisation efforts, enabling CARICOM and similar blocs to develop coherent AI governance architectures aligned with international norms while preserving sovereignty and contextual relevance 3. Promoting action-oriented mechanisms for capacity-building and resource mobilization: This includes technical assistance, funding pathways, and knowledge-sharing platforms to support AI readiness, particularly in data governance, infrastructure, and workforce development Ultimately, the Dialogue should move beyond high-level principles toward practical cooperation mechanisms, ensuring that global AI governance is not only inclusive in participation, but also equitable in outcomes

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

These priorities reflect the Caribbean's current AI development context and the need to move from fragmented, reactive adoption toward harmonised, proactive and responsible governance. Safe, secure and trustworthy AI is critical for the region, given concerns surrounding data sovereignty, dependence on external digital infrastructure, cybersecurity exposure and the use of AI systems that are neither trained on, nor designed with, Caribbean realities in mind. AI capacity-building is equally urgent, as many countries continue to face gaps in skills, infrastructure and institutional readiness, compounded by persistent brain drain and limited local AI expertise. We have also prioritised the social, economic, ethical, cultural, linguistic and technical implications of AI because these dimensions capture what makes the Caribbean context distinct, including its languages, dialects, cultural norms, small economies and social vulnerabilities. This focus shifts attention from AI as mere technology adoption to AI as a deliberate lever for inclusive growth, job creation, SME competitiveness and innovation in tourism, agriculture, finance, education, health, public services and the creative industries. Interoperability of governance approaches is also important, as CARICOM Member States would benefit from common standards and coordinated policy frameworks that reduce fragmentation, lower compliance burdens and strengthen the region's collective bargaining power with global technology providers. Finally, transparency, accountability and human oversight are essential to building public trust, reducing bias and ensuring that ethical, social and technical concerns are embedded from the outset. Taken together, these priorities would help the Caribbean, and SIDS more broadly, pursue AI in a way that is coordinated, sovereign, inclusive, economically enabling and development-focused.

A key cross-cutting issue not explicitly captured is data sovereignty and digital dependency. Caribbean states rely heavily on external cloud infrastructure and routing of data traffic, creating vulnerabilities in data jurisdiction, security, and economic value capture. AI governance should explicitly address who owns, controls, and benefits from data, particularly for developing regions/SIDS. Another emerging issue is the "connectivity paradox", where high digital adoption coexists with fragile infrastructure, high costs, and limited computational capacity. This constrains AI deployment in developing countries, regardless of user readiness levels. Thirdly, climate vulnerability and disaster risk must be integrated into AI governance discussions. For SIDS, AI is not only an economic tool but also a critical enabler of resilience, particularly in early warning systems and disaster response. Finally, brain drain versus brain circulation is an emerging dynamic. While talent loss remains a major constraint, global digital work trends create opportunities to engage diaspora expertise in AI ecosystems. These issues highlight the need for governance frameworks that are development-oriented, resilience-focused, and sensitive to structural vulnerabilities of SIDS.

Governance gaps significantly constrain the Caribbean's ability to harness AI effectively. The absence of a harmonised regional AI framework leads to fragmented policies, inconsistent safeguards, and higher compliance costs for businesses operating across multiple jurisdictions. This fragmentation weakens the region's negotiating power with global technology providers and increases the risk of vendor lock-in. Additionally, gaps in data governance and infrastructure expose the region to risks related to data sovereignty, privacy, and external dependency. Much of the region's data is stored or processed outside its jurisdiction, limiting control and value capture. Limited capacity and institutional readiness further hinder implementation. Many organisations remain in early stages of digital transformation, with insufficient skills to deploy or regulate AI systems effectively. Key Challenges: • Fragmented regulatory environment across CARICOM • Limited access to financing for AI infrastructure and innovation • Skills shortages and high rates of emigration among skilled professionals • Dependence on external digital infrastructure and cloud services • Low levels of AI industry development (less than 2% of regional AI firms in LAC) Opportunities: • Strong digital adoption (approximately 75% internet penetration in Caribbean SIDS) • High-impact use cases in priority sectors such as disaster risk reduction, agriculture, tourism, and public services • Potential for regional collaboration to achieve economies of scale • Opportunities to develop Caribbean-specific AI solutions, including language models and culturally relevant applications • Increasing global attention to inclusive AI governance, creating space for Caribbean leadership

The AI Dialogue can serve as a bridging mechanism between global best practice and regional realities. For the Caribbean, it can institutionalize mechanisms to enable intergovernmental organizations such as CTU and CARICOM to drive policy harmonization, and actively contribute to the co-development of global governance frameworks. Specific areas of action include: • Facilitating knowledge exchange and technical cooperation • Supporting alignment of regional frameworks with global standards • Providing a platform for collective advocacy by SIDS and developing regions • Enabling access to financing and partnerships for AI development

The AI Dialogue should build on national and regional initiatives in the Caribbean and, by extension, SIDS, which include: • UNESCO AI Policy Roadmap for the Caribbean • OAS Inter-American Framework on Data Governance and AI (MIGDIA) • ECLAC Digital Agenda (eLAC2026) • National AI initiatives (e.g., Jamaica AI Task Force) • CTU's Caribbean AI Task Force The interim CTU CAITF report can be found here: https://ctu.int/wp-content/uploads/2025/12/CCAITF-Draft-Interim-Report-MASTER.pdf

Embed inclusive, collaborative processes in the formulation and oversight of AI initiatives. Building on the principle of regional and multi-stakeholder cooperation, the Caribbean AI Task Force recommends formalizing channels for ongoing dialogue on AI policy. This could take the form of annual/biannual regional AI fora, coordinated by intergovernmental bodies (regional and internationally), where policymakers, industry leaders, academia, and civil society representatives review progress, share insights, and provide updated recommendations to advance AI. At the national level, governments should establish national AI advisory committees or "sandbox" regimes that include private sector and academia in pilot projects, ensuring policies remain well-informed and practical. Finally, there is need for ongoing, active engagements among national and regional stakeholders (particularly in under-represented regions and SIDS) at international fora, to under-represented perspectives and learn from global best practices, while contextualizing them to regional realities.

In the Caribbean, and by extension SIDS, under-represented perspectives in global AI governance discussions include youth, rural and underserved communities, women and girls, persons with disabilities, fisherfolk, farmers, creative workers, MSMEs, educators, public servants and disaster management practitioners and beyond. These groups are directly affected by AI-enabled public services, education systems, labour market changes, cultural production, agriculture, tourism and early warning systems, yet they often lack the resources or access to participate in global policy processes. Their inclusion requires more than open invitations; the AI Dialogue should create dedicated regional and SIDS tracks, provide travel support for participation from SIDS, support hybrid and multilingual consultations, and work through trusted regional institutions such as the CTU, CARICOM, universities, civil society networks and youth platforms. It should also include bottom-up, community-level consultations and sector-specific engagements, especially in areas of climate resilience, agriculture, education, public services and the creative industries. Together, these inclusivity strategies will support stronger, holistic AI governance, that is strongly grounded in the needs of all, and can drive meaningful and fit-for-purpose development worldwide.

For the Caribbean and other SIDS, the most effective engagement formats may be those that translate the AI Dialogue's statements into practical, multi-stakeholder action. Caribbean perspectives stress that AI governance must be harmonized across common contexts, inclusive and grounded in lived realities, particularly given the region's cultural context and vulnerabilities related to climate risk, data sovereignty, infrastructure gaps and brain drain. Useful formats could include support for, and co-development of, regional SIDS policy labs, AI governance sandboxes, youth consultations, university-based innovation labs, community dialogues and sector-specific simulations. These could focus on practical use cases such as AI-enabled early warning systems, public service chatbots, agricultural advisory tools, tourism platforms, education technologies and creative industries. Such formats would allow stakeholders to test how principles such as transparency, accountability, human oversight, data protection and inclusion apply in real Caribbean and SIDS contexts. The Dialogue should also connect global discussions with existing internet governance pathways, including the Caribbean Internet Governance Forum, the Caribbean Youth Internet Governance Forum and the wider UN Internet Governance Forum ecosystem. These spaces already provide inclusive, multi-stakeholder channels through which governments, regulators, the technical community, academia, civil society, youth and the private sector can examine digital policy issues. Leveraging these pathways would help avoid creating parallel processes and would embed AI governance within trusted regional and global digital governance mechanisms. The Dialogue should further connect with the proposed Caribbean AI Forum, national AI advisory committees and research or innovation hubs. This would allow Caribbean and SIDS perspectives to inform global AI governance, while also helping translate international outcomes into practical guidance that small states can implement. Hybrid participation, funded representation and accessible digital consultation tools would be essential to ensure that engagement is meaningful rather than symbolic.

From the Caribbean perspective, effective AI governance should be built through practical, regional mechanisms that support SIDS in overcoming limited capacity, fragmented/dated policies and regulations, and dependence on external digital infrastructure. Firstly, a harmonized Caribbean AI governance framework, such as that which is proposed by the Caribbean AI Task Force in 2026, could support the establishment of common principles, model legislation, procurement standards, risk assessment approaches, transparency requirements and audit mechanisms across the region. This would help reduce regulatory fragmentation, strengthen public trust and improve the region's collective bargaining power with global AI providers. Secondly, a Caribbean Data Commons could support responsible data-sharing in priority areas such as climate, agriculture, health, education and public administration, while protecting privacy, security and data sovereignty. This would help ensure that Caribbean data is used in ways that benefit Caribbean people and institutions. Thirdly, AI governance sandboxes and innovation labs could allow governments, universities, businesses and civil society to test AI tools in controlled environments before wider deployment. These would be especially useful for public service chatbots, early warning systems, agricultural advisory tools, tourism platforms and education technologies. Fourthly, regional capacity-building platforms, including AI education councils, research hubs and university-based innovation networks, could strengthen skills, support curriculum reform and help retain or reconnect Caribbean talent. And finally, multi-stakeholder platforms such as the IGF, regional AI fora and national AI advisory committees, can provide trusted spaces for inclusive dialogue, policy review and practical follow-up. Together, these approaches would support AI governance that is harmonized, rights-based, innovation-friendly and grounded in the realities of small states.