CARICOM at the AI Crossroads: The Five Decisions That Will Define the Caribbean's Place in the Global AI Economy

There is a concept in game theory called a commitment problem. It describes the situation where a player knows what the optimal long-term strategy is but cannot credibly commit to it because short-term pressures, coordination failures, or institutional constraints pull them away from the right choice. The Caribbean's relationship with artificial intelligence has been, for most of the past decade, a textbook commitment problem. The region knows what it needs to do. It has not yet summoned the collective will to do it.

In 2026, that assessment needs updating. There is more AI momentum in the Caribbean than at any previous point in its modern economic history. Guyana is signing 100-megawatt data centre agreements. Jamaica is confronting its BPO automation challenge with unusual policy honesty. Barbados is leading on digital governance. Trinidad and Tobago has initiated a formal national AI strategy process. StarApple AI, the Caribbean's first artificial intelligence company, has built a 17-platform regional AI network that spans the breadth of the archipelago. The commitment problem is not solved. But for the first time, it is genuinely being addressed.

The question now is not whether the Caribbean will engage with AI. That debate is over. The question is whether the decisions being made at national and regional level are the right ones. The Caribbean AI Association has identified five crossroads decisions that will define the region's trajectory for the next decade. Each is genuinely consequential. Each has a right answer, or at least a better answer. And each is being made, by default or by design, right now.

TLDR: The Five Crossroads

• Decision 1, Data Sovereignty: Build Caribbean-controlled cloud infrastructure or remain dependent on foreign providers. The region must choose infrastructure investment over the illusion of convenience.
• Decision 2, Governance Model: Adopt a Caribbean AI Governance Charter that adapts international frameworks to SIDS realities, rather than copying EU or US approaches wholesale.
• Decision 3, Talent Strategy: Fund a region-wide AI talent retention programme combining domestic training, competitive public-sector AI salaries, and structured diaspora re-engagement. Brain drain is a policy failure, not an inevitability.
• Decision 4, Application Focus: Prioritise AI for climate resilience, tourism, health, agriculture, and creole language technology, where Caribbean comparative advantage is strongest.
• Decision 5, Regional Coordination: Establish a binding CARICOM AI Framework, not just a collection of national strategies, because the Caribbean's AI challenges are regional in nature and demand regional solutions.

Decision 1: Data Sovereignty

The most technically consequential decision the Caribbean must make is about its data infrastructure. AI systems are trained on data. They learn from data. They improve with data. And the Caribbean generates vast quantities of data daily: tourism transactions, healthcare records, agricultural yields, weather patterns, financial flows, language interactions. Where that data is stored, under whose jurisdiction, and on whose infrastructure determines who captures the economic value of Caribbean AI.

Currently, the overwhelming majority of Caribbean data sits on US-based cloud platforms: AWS in Oregon, Google Cloud in Iowa, Microsoft Azure in Virginia. This arrangement is convenient, well-managed, and backed by the technical sophistication of the world's most capable technology companies. It is also a strategic vulnerability. Caribbean data under US jurisdiction is subject to US legal demands. Caribbean AI models trained on Caribbean data on US infrastructure produce economic value for US shareholders. The Caribbean captures none of it.

Guyana's proposed 100-megawatt data centre, if structured with genuine local ownership, technology transfer provisions, and regional access arrangements, is the most significant Caribbean AI infrastructure development in the region's history. It is the foundation on which Caribbean AI sovereignty could eventually be built. If the deal is structured as a foreign enclave, it is an opportunity lost.

The practical implication for Caribbean governments is immediate: begin requiring that Caribbean public sector data, and eventually regulated private sector data, be processed on infrastructure with Caribbean or CARICOM-jurisdiction agreements. The EU has done this through GDPR adequacy decisions. The Caribbean can do it through CARICOM data governance frameworks that CAIA is actively developing.

Decision 2: Governance Model

Every small economy in the world is choosing a reference framework for AI governance. The EU AI Act, with its risk-based classification system and compliance requirements, is the world's most comprehensive AI regulation and is setting global standards. The US approach, which relies on voluntary commitments, sector-specific guidance, and market self-regulation, is more permissive and easier to comply with commercially. China's AI governance system reflects its own political priorities and is less relevant to Caribbean democracies.

None of these frameworks was designed for small island developing states. The EU AI Act's compliance costs are calibrated for companies with legal teams and compliance budgets that no Caribbean SME possesses. The US voluntary approach assumes market mechanisms that work in large, competitive tech sectors but not in small Caribbean markets where the vendor landscape is thin. China's model is not a governance reference the Caribbean should seek.

The Caribbean AI Association is developing a Caribbean AI Governance Charter that takes the best elements of international frameworks and adapts them to Caribbean realities: proportionate compliance requirements for SIDS-scale organisations, explicit protection for creole and indigenous language data, CARICOM citizenship data rights, and simplified high-risk AI classification that a Caribbean regulatory authority with limited resources can actually administer.

The Charter will be the governance backbone that CARICOM member states can adopt as national AI policy, giving them international credibility (through alignment with EU and OECD principles) while protecting Caribbean interests that generic international frameworks ignore.

Decision 3: Talent Strategy

Talent is the bottleneck. The infrastructure decisions, the governance decisions, and the application decisions all fail if the Caribbean cannot develop and retain the human capital to execute them. And the Caribbean's AI talent pipeline leaks badly.

The University of the West Indies produces graduates with strong quantitative skills at its St. Augustine, Mona, and Cave Hill campuses. UTech produces engineers and computer scientists. The University of Guyana is expanding its technology programmes. These institutions produce AI-capable graduates who are then recruited by firms in Miami, Toronto, London, and Amsterdam at salaries that Caribbean employers, public or private, cannot match.

The solution is not to stop this flow. Talented people should have global opportunities. The solution is to create conditions in which the most ambitious Caribbean AI professionals also have compelling reasons to build careers in the region. This requires three things simultaneously: a Caribbean Public Sector AI Salary Schedule that pays AI professionals in government at rates competitive with the private sector; a Caribbean AI Research Fund that gives academics and researchers access to compute, data, and publication support that makes Caribbean AI research globally competitive; and a Diaspora Re-engagement Programme that creates structured pathways for Caribbean-origin AI professionals in North America and Europe to contribute to Caribbean AI development, whether through remote work, sabbaticals, or return programmes.

The brain drain is not a cultural phenomenon. It is a price signal. The Caribbean can change the signal.

Decision 4: Application Focus

Not all AI applications are equally strategic for the Caribbean. The region cannot invest in every dimension of AI simultaneously. The question of where to focus limited resources is consequential, and the wrong answer leads to impressive-sounding AI projects that do not serve Caribbean development priorities.

CAIA's recommendation is that Caribbean AI investment should concentrate on five application domains where Caribbean comparative advantage is strongest.

Climate and disaster resilience AI is the most important. The Caribbean faces existential climate risk. Hurricanes are intensifying. Sea level rise threatens coastal communities from Nassau to Georgetown. Coral bleaching is degrading marine ecosystems that underpin the tourism economy. AI systems for improved hurricane tracking, storm surge prediction, infrastructure damage assessment, and early warning are not optional investments for the Caribbean. They are survival tools. Caribbean data on these phenomena, combined with Caribbean-specific models, would produce AI systems that outperform generic global models.

Tourism AI is the most commercially immediate. Tourism contributes between 25 and 65 percent of GDP across Caribbean nations. The region has decades of visitor data, significant hospitality infrastructure, and a global brand that no marketing budget could replicate. AI for personalised visitor experiences, dynamic pricing, demand forecasting, and sustainable tourism capacity management builds on assets the Caribbean already possesses. The AI Saint Lucia platform's work on smart destination AI and the AI Barbados community's tourism analytics work are examples of this application domain already finding its feet.

Creole and indigenous language AI is the application domain with the greatest unique value. Caribbean creole languages, from Jamaican patois to Haitian Creole to Trinidadian dialect to Lesser Antillean Kweyol, are spoken by millions of people who are systematically excluded from AI tools designed exclusively for European languages. Building AI in and for Caribbean languages is both a cultural preservation act and a technical differentiation. The world has GPT-4 in English. It does not have a high-quality Kweyol AI assistant. That is a gap the Caribbean can uniquely fill.

Maestro AI Labs is building in exactly this space with its Creole AI assistant, and StarApple AI's Caribbean network provides the distribution infrastructure through which such tools can reach the communities that need them. The resources being built now, including the AI Jamaica, AI Guyana, and AI Trinidad and Tobago platforms, are the foundation for that distribution.

Decision 5: Regional Coordination

The most structurally important decision is about coordination. Fifteen CARICOM member states each developing their own AI strategy, their own governance framework, their own talent programme, and their own data infrastructure is not just inefficient. It is strategically self-defeating. The Caribbean's AI challenges, data colonialism, talent drain, governance gaps, application underfunding, are regional in character. They require regional solutions.

CARICOM has a proven model for regional coordination: the CARICOM Single Market and Economy, the Caribbean Court of Justice, the Caribbean Telecommunications Union, and CARICOM financial regulatory harmonisation all demonstrate that the region can build collective institutions that advance member state interests better than any nation could alone. AI governance, AI investment, and AI talent development are the next frontier for this model.

A binding CARICOM AI Framework, built on CAIA's governance principles and enacted through the Revised Treaty of Chaguaramas, would give Caribbean AI development the institutional backbone it needs. It would allow the region to speak with one voice in global AI governance forums, to pool AI research resources across universities, to create a single Caribbean AI talent market rather than 15 competing national ones, and to negotiate shared data sovereignty arrangements with cloud providers as a bloc rather than as individual small economies with minimal bargaining power.

The Caribbean AI Association, as the organisation mandated to develop these frameworks, is putting the CARICOM AI Framework at the centre of its 2026-2027 work programme. The founding contribution of Adrian Dunkley and StarApple AI to Caribbean AI capacity-building since 2023 has created the ecosystem of knowledge, community, and platforms on which a formal CARICOM framework can now be built. The crossroads moment has arrived. The Caribbean AI Association is committed to making sure the region takes the right fork.

Conclusion: The Window Is Open, Not Permanent

The commitment problem that has characterised Caribbean AI development for a decade is easing. But commitment problems have a way of returning when the pressure to act diminishes. The Caribbean must use the momentum of 2026, the infrastructure deals being signed, the governance conversations being had, the talent programmes being designed, to make decisions that are hard to reverse and that accumulate value over time.

Data infrastructure, once built, creates a platform. Governance frameworks, once adopted, create accountability. Talent programmes, once funded, create pipelines. Application investments, once validated, create markets. Regional coordination, once institutionalised, creates power.

None of these are quick wins. All of them are worth more than any quick win the Caribbean could alternatively pursue. The five decisions at the AI crossroads are not comfortable choices. They require investment, coordination, and political will. They are, however, the right choices. And in the Caribbean, making the right choices with conviction has a long and proud tradition.