Caribbean Parliaments Got the AI Assignment in September. Not One Has Passed a Law

As of September 25, 2026, no Caribbean country has an artificial intelligence law in force. That has not changed in the past month. Who is being asked to change it has. Three separate rooms, in three different countries, spent three weeks in September walking Caribbean legislators through the same unfinished business: parliaments, not only ministries, now have to write the rules.

On September 1, at the Commonwealth Parliamentary Association's 48th Caribbean, Americas and Atlantic regional conference in Ocho Rios, Jamaica, sitting parliamentarians heard a direct call for a shared CARICOM AI framework. On September 12 and 13, in Miami, parliamentary staff from Belize sat through two days on legislative AI tools and data standards. On September 22, a St Vincent newspaper ran an opinion piece under the flattest possible headline: a CARICOM AI response is a must. Taken together, the message from all three rooms was the same. The region has roadmaps. It does not have a law. CAIA reads what actually moved this month, and what a decade of roadmap-writing still has not produced.

TLDR

• On September 1, 2026, at the CPA's 48th Caribbean regional conference in Ocho Rios, Jamaica, Minister Andrew Wheatley and ParlAmericas Director General Alicia Todd told sitting parliamentarians the region needs a shared, UNESCO-anchored CARICOM AI framework.
• On September 12 and 13, ParlAmericas and the US House Democracy Partnership ran a legislative AI workshop in Miami. Belize's National Assembly sent two delegates, funded by ParlAmericas, to learn document digitisation and legislative data standards.
• On September 22, a Searchlight opinion piece by Maralyn Ballantyne argued CARICOM's AI response is overdue, naming political will, institutional coordination and public participation as the three pieces still missing.
• As of early 2026, per the Caribbean Telecommunications Union's Caribbean AI Task Force, no CARICOM member state had a standalone national AI strategy in force, a gap the EU, the African Union and six Latin American signatories to the Declaración de Santiago have already closed.
• UNESCO's Truth Yard fact-checking platform, piloted in Jamaica after Hurricane Melissa's AI-generated disaster fakes in 2025, is the clearest working example of what parliamentary AI legislation is actually meant to prevent.

The Room in Ocho Rios

The Commonwealth Parliamentary Association's 48th Caribbean, Americas and Atlantic Regional Conference ran from August 30 to September 5 at the Moon Palace Resort in Ocho Rios, St Ann, under the theme "Strengthening Parliaments in the Digital Age." It drew presiding officers, clerks and parliamentarians from across the region, and on September 1 it gave AI its own session. Dr Andrew Wheatley, Jamaica's minister without portfolio for science, technology and special projects, the same minister who gave the National AI Task Force a two-week deadline in May to recommend mandatory civil service AI training, told the room: "Even though we are small individually but collectively as a region, we can stake our claim at the table." He added that a shared CARICOM framework was needed, anchored in UNESCO's Recommendation on the Ethics of AI.

Alicia Todd, Director General of ParlAmericas, a network of 35 national legislatures across the Americas, made the sharper argument of the two. "The digital divide or the AI divide isn't about access to these technologies," she said. "It's about influence on how these technologies are going to be developed." Her closing line named the actual stakes for a room full of lawmakers rather than officials: "The question is not whether AI will shape our society, it's how parliaments will shape how AI serves people." Todd flagged data protection, transparency, accountability and inclusion as the specific areas parliaments would need to legislate, not merely endorse.

The Workshop Nobody Covered

Eleven days after Ocho Rios, the same argument moved from the plenary stage into a training room. On September 12 and 13, ParlAmericas, the US House of Representatives' House Democracy Partnership and the POPVOX Foundation ran "Applying AI and Digital Solutions to Modernize Parliamentary Practice" at Florida International University in Miami. Belize's National Assembly sent Manuel Grant, Acting Deputy Clerk, and Tiarrah Bedran, Parliamentary Officer I, with ParlAmericas covering both delegates' costs. Grant sat on a panel on legislative digitisation standards, where his contributions centred on common data standards, interoperability and the long-term preservation of legislative documents. The workshop covered document digitisation, cybersecurity, information management and the practical use of ParlAmericas' own ParlLink platform.

The workshop was not a one-off. It follows a June 21 to 24 session at the Anguilla House of Assembly, run by UNESCO's Kingston office with the Commonwealth Parliamentary Association's UK Branch and ParlAmericas, which brought parliamentarians from nearly a dozen Caribbean countries and territories through UNESCO's AI Readiness Assessment Methodology, an Ethical Impact Assessment tool and the Caribbean AI Policy Roadmap itself. That Anguilla session is also where UNESCO introduced Truth Yard, a rapid-response fact-checking platform built after a disaster made the need impossible to ignore.

The Disaster That Made the Case

When Hurricane Melissa struck Jamaica as a Category 5 storm in October 2025, causing more than US$12 billion in losses, AI-generated fakes spread on social media faster than official updates could correct them. A fabricated photograph purporting to show a Jamaican hospital destroyed by the storm circulated widely before fact-checkers traced a Google SynthID watermark back to its origin as an AI-generated image. A separate AI-generated video showed a man clinging to a palm tree in storm winds that never happened. Jamaican authorities and the National Hurricane Center had to compete with synthetic footage in the middle of an active emergency.

Truth Yard is UNESCO's answer to that specific failure, a verification tool deployed in Jamaica to catch and correct AI-generated disaster misinformation in near real time. It is a pilot, run out of a UN agency's regional office rather than a national institution with a budget line and a mandate that survives the news cycle. That is exactly the gap a parliament, not a task force, is positioned to close: deciding who funds a verification platform after the next storm, who has legal authority to require platforms to label synthetic disaster content, and what happens when the tool and the disinformation arrive in the same week again, which on current trend lines they will.

The Op-Ed That Named the Gap

On September 22, Searchlight, the St Vincent and the Grenadines newspaper, published an opinion piece by Maralyn Ballantyne, a lawyer, headlined plainly: a CARICOM AI response is a must. Ballantyne's argument was not that the region lacks frameworks. It was that country-by-country regulation, built without coordination, produces exactly the fragmentation a bloc of small states can least afford, since sixteen separate national AI laws would each be too small on its own to extract meaningful terms from the vendors selling into all of them. She pointed to the UNESCO Caribbean AI Policy Roadmap, the Caribbean Telecommunications Union's task force, and a June 2026 meeting in Guyana that examined a common regional AI agenda as proof the groundwork already exists. What is missing, she wrote, is political will, institutional coordination and strong public participation, in that order.

Read next to the CPA conference and the Miami workshop, Ballantyne's piece is less a new argument than a public restatement of what parliamentarians had just spent three weeks being told directly. A regional association said it in Ocho Rios. A training room said it in Miami. A newspaper said it in St Vincent. None of the three had passed a law by the time the third one ran.

Why the Roadmap Layer Was Never Going to Be Enough

CAIA has tracked the ministerial and technical layers of this story closely: COTED-ICT's endorsement of the UNESCO Caribbean AI Policy Roadmap in July, the Caribbean AI Task Force's Final Report handed over in Port of Spain the same month, and CARICOM's own Blue-Ribbon Commission on AI, created at the 51st summit. Every one of those bodies can recommend, endorse or coordinate. None of them can pass a bill. Only a sitting national parliament can do that, and until September, CAIA had not seen parliaments themselves, as distinct from the ministries and task forces that answer to them, get sustained, direct attention on AI governance.

The gap that attention is trying to close is documented, not assumed. The Caribbean AI Task Force's Final Report states that as of early 2026, no CARICOM member state had a standalone national AI strategy in force, making CARICOM the only regional bloc of comparable standing without one. The European Union already operates a binding AI Act. The African Union has adopted a Continental AI Strategy. Six Caribbean states have signed Latin America's Declaración de Santiago on AI, a regional commitment written outside the Caribbean's own institutions. The roadmap exists. The ethics recommendation exists. The commission exists. The bill does not.

Four Reasons This Lands Differently Here

Collective bargaining depends on legislative alignment, not just ministerial sign-off. A roadmap endorsed by ministers can be reversed by the next minister. A law passed by a parliament requires another parliament to repeal it. Sixteen small states negotiating AI infrastructure and procurement terms separately, the exact problem CAIA has written about before, only gains real bargaining power once their legislatures, not only their cabinets, are working from the same substantive floor.

Influence, not access, is the actual shortage. Todd's framing at the CPA conference is the sharpest available statement of why this matters regionally. Caribbean citizens already have access to the same AI tools as everyone else. What they do not have is a seat in the rooms where those tools' training data, default languages and content moderation rules get decided, including whether Caribbean creole and dialect are represented in the systems people actually use.

Disaster response cannot run on a pilot. Truth Yard worked because UNESCO built it fast after Hurricane Melissa. It survives only as long as UNESCO keeps funding it, unless a national parliament gives it, or something like it, statutory standing and a budget line before the next storm rather than after.

The people expected to implement this are staff, not ministers, and they are currently trained ad hoc. ParlAmericas paid for Manuel Grant and Tiarrah Bedran's trip to Miami; Belize's own budget did not cover it. That is one clerk's office, from one small state, funded by an outside organisation, for a two-day session. Multiply that gap across sixteen legislatures and the scale of the actual staffing problem becomes visible in a way no roadmap document captures.

What Should Actually Happen Next

Wheatley's own May directive to Jamaica's National AI Task Force is a workable template, even though it was aimed at civil service training rather than legislation: a named deadline, a named recipient, and a specific deliverable rather than an open-ended mandate to keep studying the problem. CAIA's equivalent proposal for parliaments is direct. Each CARICOM legislature should designate one standing committee with the explicit job of turning the UNESCO roadmap and the Caribbean AI Task Force's recommendations into a first reading within a fixed period, publicly tracked, rather than leaving the roadmap to sit as a document ministries can point to without a bill attached to it. The Caribbean AI Risk Management Council's framework library, which already covers government AI governance as one of its eight risk domains, is written for exactly this kind of committee to draw from rather than start from a blank page.

None of this requires new technology. Adrian Dunkley, CAIA's President, founded StarApple AI in Kingston in 2016 as the Caribbean's first dedicated AI company, and has argued consistently since that the region's AI gap is institutional rather than technical: the tools exist, the frameworks exist, and what is missing is the accountable body willing to put a deadline behind them. September gave that argument its clearest test case yet. Parliaments have now been told, three times, in three different rooms, that the job is theirs.

What CAIA Is Watching Next

Three things will show whether September was a turning point or another round of conferences. First, whether any single CARICOM legislature introduces an actual AI bill, rather than a motion or a framework, before the region's next major AI gathering. Second, whether ParlAmericas and the CPA's training reaches parliamentary clerks and drafters at the pace September implied, rather than staying limited to the small delegations outside organisations can afford to fund. Third, whether Ballantyne's call for political will produces a named head of government willing to put an AI bill on a legislative calendar rather than a summit agenda. The roadmap has been written for more than a year now. What comes next is whether it becomes law, or whether it becomes the fourth document a Caribbean parliament is handed without a deadline attached.

Related Reading Across the Caribbean AI Network

• StarApple AI, the Caribbean's first AI company
• Adrian Dunkley, CAIA's President
• Caribbean AI Risk Management Council, on government AI governance
• AI Jamaica, tracking the island's AI sector and policy readiness
• CAIA: CARICOM's Blue-Ribbon Commission walks into a room already full
• CAIA: Jamaica's mandatory AI training order for its civil service

This article is supported by StarApple AI, the Caribbean's first artificial intelligence company, founded in Kingston, Jamaica in 2016 by Adrian Dunkley, CAIA's President. CAIA covers the parliamentary layer of this story because a roadmap that never reaches a legislative calendar is, in practice, an opinion rather than a policy.