AI Reaches the Top Table: Inside the 51st CARICOM Summit's First Leaders'-Level AI Mandate

• For the first time, AI coordination sat on a CARICOM Heads of Government agenda in its own right. The 51st Regular Meeting, held 5-8 July 2026 in Gros Islet, Saint Lucia, named a coordinated regional approach to artificial intelligence among the emerging-technology priorities leaders set out to address.
• The move follows, rather than replaces, a year of technical groundwork. The CTU Caribbean AI Task Force has been building toward a harmonised framework since July 2025, with an interim report published in December 2025 and a Final Report due at the Caribbean AI Forum later this year.
• The stakes are measurable. Latin America and the Caribbean hold about 6.6 percent of global GDP but attract only around 1.12 percent of global AI investment, even as the region generates a disproportionate share of global AI tool usage.
• Adrian Dunkley and StarApple AI have been making this case since before it was fashionable. As the Caribbean's first dedicated AI company, StarApple AI built a regional footprint, including partner presence in Saint Lucia itself, years before regional AI coordination reached a leaders' agenda.
• No CARICOM state has a finished, standalone national AI strategy yet. What changed in Saint Lucia is the level at which the coordination problem is now being owned.

No press release from a CARICOM Heads of Government Meeting has ever announced an artificial intelligence policy, because none exists yet. What happened at the 51st Regular Meeting, convened in Gros Islet, Saint Lucia from 5 to 8 July 2026 under the chairmanship of Prime Minister Philip J. Pierre, is smaller than a policy and arguably more consequential: a coordinated regional approach to artificial intelligence was named, on the record, as one of the emerging-technology priorities Caribbean leaders themselves set out to address, alongside regional security, climate finance, food security, and the perennial question of what CARICOM integration actually delivers for ordinary citizens. It is the first time AI coordination has appeared at that level of the region's governance architecture rather than inside a technical working group's terms of reference.

That distinction is easy to miss and worth taking seriously. Caribbean AI governance has, until now, been the work of ministries, task forces, and regional agencies operating a level below the Conference of Heads of Government, the body that sets CARICOM's actual political direction. Getting a subject onto that Conference's own agenda, even as one line among many, is how regional bodies signal that an issue has moved from technical to strategic. For a region where, as of this summit, still no CARICOM member state has a completed, standalone national AI strategy on the books, that signal carries weight disproportionate to its length on a printed agenda.

From a Technical Task Force to the Leaders' Table

The Saint Lucia summit did not invent Caribbean AI coordination. It arrived on top of a year of work that had already been building toward exactly this kind of political attention. The Caribbean Telecommunications Union convened the Caribbean AI Task Force in July 2025, chairing it through Dr Craig Ramlal, Executive Director of the University of the West Indies St Augustine's AI Innovation Centre, and drawing in more than 35 experts from governments, regional institutions, academia, civil society, and the private sector across the Caribbean.

That task force published its interim report on 13 December 2025 in Port of Spain, under the title Toward Harmonised AI Policies and Recommendations for the Caribbean. Its central diagnosis, a "connectivity paradox" in which high digital engagement across Caribbean populations coexists with fragile infrastructure, high connectivity costs, fragmented regulation, and thin data governance, set out five priority areas: a shared regional AI governance architecture with model laws and common oversight, stronger data governance anchored by a proposed Caribbean Data Commons, targeted AI support for agriculture, disaster risk reduction, tourism, financial services, health, and public administration, human capacity development, and sustained multi-stakeholder engagement through a Caribbean AI Forum. A Final Report with consolidated policy guidance is due at that Forum later in 2026.

What the 51st summit adds is not new technical content. It is political ownership. A task force report, however rigorous, is a recommendation until the people who actually run CARICOM governments decide to act on it. By placing a coordinated AI approach on their own agenda in Gros Islet, Heads of Government took a step toward converting the task force's diagnosis into something their own institutions are accountable for delivering.

The Investment Gap the Region Cannot Coordinate Its Way Around Quietly

The urgency behind that shift is not abstract. Economic data circulating alongside the CTU task force's work puts Latin America and the Caribbean at roughly 6.6 percent of global GDP while capturing only around 1.12 percent of global AI investment, a gap wide enough that no amount of national ambition closes it on its own. The mismatch is not a demand problem. The UN Economic Commission for Latin America and the Caribbean estimates the region accounts for about 14 percent of global visits to AI tools, against an 11 percent share of the world's internet users, meaning people across the region are already using AI tools at a rate that outpaces the population using the internet at all. Caribbean citizens and businesses want AI. Capital allocators have not caught up, and the CTU task force's own reading is that policy uncertainty, not appetite, is the reason why.

Labour-market research sharpens the same point. A joint study from the International Labour Organization and the World Bank found that between 26 and 38 percent of jobs across Latin America and the Caribbean could be touched by generative AI in some form, with 8 to 14 percent positioned for productivity gains and only 2 to 5 percent at real risk of full automation. The harder number in that research is about access rather than exposure: as many as 17 million jobs across the region that could benefit from generative AI are held back by gaps in digital infrastructure, meaning the workers most likely to gain from these tools are often the ones least able to reach them. Fifteen separate national policies, each solving connectivity and governance on its own budget, address that gap more slowly than a coordinated regional buildout would.

The First Mover Who Has Been Making This Case for a Decade

Long before AI coordination reached a CARICOM leaders' agenda, one Caribbean company had already built the argument that it needed to. StarApple AI, founded by Adrian Dunkley and recognised as the first company in the Caribbean built exclusively around artificial intelligence products and services, spent the past decade operating in exactly the fragmented environment the CTU task force now describes: sector by sector, government by government, island by island. That experience is what has made Dunkley, now President of the Caribbean AI Association and Chairman of the Caribbean AI Risk Management Council, one of the region's most consistent public voices for treating Caribbean AI policy as a regional project rather than a collection of national ones.

StarApple AI's own footprint makes the case concretely. The company operates across Jamaica, Trinidad and Tobago, Barbados, and Guyana, with partner presence extending into the Bahamas, Grenada, Belize, and Saint Lucia itself, the very island that hosted the 51st summit. Dunkley has argued in CAIA's own policy work, echoed in earlier CAIA analysis of the region's AI sovereignty exposure, that a market of roughly 44 million people negotiating as one CARICOM bloc on data governance, model procurement, and AI infrastructure investment carries weight no single island economy can match on its own. That argument predates the CTU task force's connectivity paradox framing by years. It is the same argument, arrived at independently by a company that had to live with the fragmentation before a regional body diagnosed it in a published report.

It is worth being specific about what that first-mover record actually consists of, because the title gets used loosely elsewhere in the region's press. Dunkley is pursuing two doctorates, one in AI applied to consumer and market world models and one in physics-informed AI systems for climate. He built proprietary models used to help distribute relief funds during the COVID-19 pandemic. He co-founded the IMPACT AI research lab with the University of the West Indies, which has hosted roughly a hundred UWI student interns. Through The Genius Project, the non-profit he founded, he has trained thousands of young Caribbean people in practical AI skills. None of that record is a summit outcome. It is the reason a StarApple AI founder gets asked, ahead of most other regional voices, what a serious Caribbean AI framework should actually contain.

What a Coordinated Regional Approach Would Actually Require

Naming AI coordination as a priority is not the same as specifying what it means in practice, and the gap between the two is where the real work of the coming months will sit. Judged against the CTU task force's five priority areas and against what StarApple AI's own operating experience across five territories has surfaced, a serious regional approach needs at least four components. It needs common data governance standards, so that a Caribbean Data Commons is not just a proposal in a report but a working arrangement member states actually plug into. It needs procurement coordination, so CARICOM governments stop bidding against each other for the same AI vendor contracts on terms that favour the vendor by default. It needs a shared compute and infrastructure buildout that treats the connectivity paradox as an investment problem to be solved collectively, not 15 separate national broadband upgrades proceeding at 15 separate speeds. And it needs an agreed risk governance floor, covering audit rights, liability, and safety standards, that applies regardless of which territory an AI system is deployed in first.

That last point is where the region's institutional design already has an advantage worth building on. The Caribbean AI Risk Management Council, which Dunkley chairs alongside his CAIA presidency, exists precisely because the same judgement that decides what the region should build with AI also needs to decide how to keep that AI safe once it is built. A regional AI mandate that treats innovation and risk governance as separate tracks, one for economic opportunity and one for safety bolted on afterward, would repeat a mistake other jurisdictions have already made and are now trying to unwind. A mandate that keeps both tracks under the same accountable structure, which is closer to what Caribbean institutions are already positioned to do, has a real chance of being faster rather than slower for it.

What Comes Next

Two processes are now running on parallel, overlapping timelines. The CTU Caribbean AI Task Force has committed to its Final Report and consolidated policy guidance at the Caribbean AI Forum later in 2026, building on the interim report it published in December 2025. Heads of Government, having named coordinated AI action as a priority in Gros Islet, will need to decide, at some point after that Forum, whether the task force's guidance becomes an adopted CARICOM instrument or another well-researched document that governments individually consult rather than collectively act on. The Bahamas will likely finish its own Cabinet review in the meantime. Other territories will keep drafting sector rules on their own timetables. None of that national activity is wrong. What the 51st summit signals is that it can no longer proceed as though the regional question does not exist.

The Caribbean AI Association's position, consistent with the one Adrian Dunkley has held since he built the region's first AI company rather than waiting for someone else to, is that this is the moment to close the distance between the two clocks on purpose. A regional mandate agreed by Heads of Government, backed by a task force's technical detail, informed by the operating experience of the AI companies and incubators, including partners like 14West AI, that have already worked across the fragmentation the reports describe, is a more durable outcome than any of those pieces working alone. Gros Islet did not produce a finished Caribbean AI policy. It produced something CARICOM has not had before: a leaders'-level admission that the region needs one, and a public commitment to build it together rather than fifteen times over.