Nineteen Parliaments, One AI Plenary: What the Caribbean's Legislatures Actually Discussed

TL;DR

• The 48th CPA Caribbean, Americas and Atlantic Regional Conference ran 30 August to 5 September 2026 at Moon Palace Jamaica in Ocho Rios, drawing roughly 70 delegates from all 19 member states plus observers from Suriname and St Martin.
• Its 1 September plenary was the region's first joint parliamentary session dedicated to AI governance, distinct from the executive-branch work already under way at CARICOM.
• Minister Dr Andrew Wheatley called for CARICOM states to negotiate AI rules as a bloc and named data protection, algorithmic transparency and Creole-language representation as priority gaps.
• ParlAmericas Director General Alisha Todd told delegates the region's AI divide is about influence over how systems are built, not access to them.
• Per the DPO Caribbean tracker, Dominica, Saint Kitts and Nevis, and Suriname still have no national AI policy document, while commercial AI is already live at scale: Digicel's Ruby handled 535,000-plus Jamaican conversations in the year to March 2026.
• The plenary raised the right questions but did not announce a mechanism connecting its own recommendations to the model legislation each of the 19 parliaments would individually need to pass.

Nineteen Caribbean, Americas and Atlantic legislatures sent their speakers, presidents and clerks to Ocho Rios, St Ann, for a week that ran 30 August to 5 September 2026, and for the first time, one full plenary session was built entirely around artificial intelligence. That single fact matters more than any specific line from the session itself: every AI governance conversation the region has had in 2026, from the CARICOM Blue-Ribbon Commission to the UNESCO policy roadmap, has run through executive-branch ministries and regional secretariats. On 1 September, it was the people who actually vote laws into force who sat in the room.

The 48th Regional Conference of the Commonwealth Parliamentary Association's Caribbean, Americas and Atlantic (CAA) Region, hosted by the Jamaican Parliament under Speaker Juliet Holness and themed Strengthening Parliaments in the Digital Age, brought together approximately 70 delegates representing all 19 CPA CAA member states, alongside observer delegations from Suriname and St Martin. Prime Minister Dr Andrew Holness delivered the keynote at the 31 August opening, and CPA Secretary-General Stephen Twigg attended alongside ParlAmericas Director General Alisha Todd. Buried among sessions on online bullying, disability representation and parliamentary decorum in the digital age sat the AI plenary, and what got said there deserves more attention than it has had.

Why a Parliamentary AI Session, and Why Now

The timing was not incidental. The conference landed five weeks after CARICOM's Council for Trade and Economic Development's ICT configuration endorsed the UNESCO Caribbean Artificial Intelligence Policy Roadmap on 7 July 2026, and four weeks after CARICOM Heads of Government agreed in Saint Lucia on 3 August 2026 to establish a Blue-Ribbon Commission on Artificial Intelligence. Both of those were executive and intergovernmental moves: a Council of ministers endorsing a roadmap, a summit of heads of government standing up an advisory commission. Neither involved a vote in a national legislature, because neither required one at that stage.

That is precisely the gap the CPA CAA plenary exists to close. A roadmap can recommend, and a commission can advise, but data protection standards, algorithmic accountability requirements and consumer AI disclosure rules only bind once a parliament passes them into national law. Nineteen separate legislatures, in nineteen separate legislative calendars, is the actual bottleneck standing between any regional AI recommendation and an enforceable rule on the ground in Kingston, Bridgetown or Georgetown. Getting the people who run those calendars into one room, even for a single session, is worth more than another communique.

What Was Actually Said in the Room

Dr Andrew Wheatley, Jamaica's Minister without Portfolio in the Office of the Prime Minister with responsibility for Science, Technology and Special Projects, framed the region's leverage problem directly: individually, CARICOM states carry little weight at the tables where global AI standards get set, but collectively, as CARICOM and as the wider Latin American and Caribbean bloc, the region can stake a claim. He named data as "an intrinsic asset that must be preserved and must be guarded," and pushed specifically on the underrepresentation of Caribbean English and Creole languages inside AI systems, a gap that is technical rather than rhetorical: a large language model trained overwhelmingly on North American and European text handles Jamaican Patois, Trinidadian Creole or Guyanese Creolese poorly, if at all, which has direct consequences for any government service that tries to deploy AI-driven citizen support in the language people actually speak.

Alisha Todd, ParlAmericas' Director General, reframed the stakes for the room in a line worth repeating in full: the AI divide in this hemisphere is not about access to the technology, it is about influence over how it gets developed, deployed and governed. That distinction is the one most global AI commentary misses when it discusses the Caribbean. Access is not the region's binding constraint; Digicel, telecoms and banks across the territories are already deploying AI tools at scale. Influence over the standards those tools operate under is what nineteen small legislatures, working separately, structurally lack.

The session's stated focus areas, data protection, privacy, transparency, accountability, gender equality and inclusion, tracked closely with the four pillars of the UNESCO Caribbean AI Policy Roadmap that COTED-ICT had endorsed weeks earlier. What the plenary did not produce, at least in the public conference materials available as of this writing, was a specific mechanism for turning that shared vocabulary into model legislation any of the 19 parliaments could adopt directly. A youth-focused companion event, the 19th Regional Youth Parliament, convened virtually on 3 September under the theme Youth Leadership in the Age of Artificial Intelligence, extending the same conversation to the delegates who will eventually inherit whatever framework, or absence of one, the current generation of legislators leaves behind.

The Institutional Gap: Legislatures Are Behind Their Own Executives

Set the CPA CAA plenary against the DPO Caribbean AI Laws and Regulations tracker and the gap between conversation and codification becomes concrete. As of this writing, Dominica, Saint Kitts and Nevis, and Suriname each show no substantive national AI document identified. Across the entire CARICOM and wider Caribbean membership, only Jamaica, with its National Artificial Intelligence Policy Recommendations spanning nine areas from workforce development to legal frameworks, and the Dominican Republic, with a published National AI Strategy, have country-level AI policy documents in place at all. No CARICOM-wide document has been formally adopted as regional law; the UNESCO Roadmap and the Blue-Ribbon Commission are both advisory, not binding.

Meanwhile, the commercial deployment of AI in the region has not waited. Digicel's AI assistant Ruby, built on a platform developed by the Canadian firm Ada, evolved during the financial year ended March 2026 from a basic knowledge tool into what the company calls an AI-powered smart digital assistant, available across WhatsApp, voice, web and the MyDigicel app. In Jamaica alone, Ruby supported more than 535,000 conversations and resolved 55 percent of them without human agent involvement, according to Digicel's 2026 Sustainability Report, and the company has confirmed plans to expand the enhanced version to additional Caribbean markets during the year ending March 2027. That is a live, consumer-facing AI system handling more than half a million interactions in a single territory that, notably, does have a national AI policy document. Extend Ruby's rollout to Dominica, Saint Kitts and Nevis, or any of the other territories the DPO Caribbean tracker flags as having no AI document at all, and the mismatch between deployment speed and legislative speed becomes the story regulators will eventually have to answer for after the fact rather than in advance of it.

This is exactly the terrain the Caribbean AI Risk Management Council was built to work: not whether AI arrives in the Caribbean, since it plainly already has, but whether the region's governance keeps pace with where it lands first. A parliamentary plenary that names data protection and algorithmic transparency as priorities is a necessary first conversation. It is not, on its own, the regulatory instrument that would have applied a consent standard to Ruby's rollout before it reached its 535,000th conversation.

A Pan-Caribbean Reading: Beyond the Communique

What happened in Ocho Rios sits inside a wider pattern of regional institutions circling the same problem from different angles without yet connecting their work. CARICOM's Blue-Ribbon Commission operates at the level of Heads of Government. COTED-ICT's endorsement of the UNESCO Roadmap operates at the level of a specialised ministerial council. The Caribbean Telecommunications Union's Caribbean AI Task Force, launched in July 2025 and delivering its final report and consolidated policy guidance in July 2026, operates at the level of technical harmonisation across telecommunications regulators. The CPA CAA plenary now adds a fourth track, at the level of the legislatures themselves. Four institutional layers, each doing real work, none of them yet formally wired to the others in a way that turns a roadmap pillar into a bill a national parliament actually reads a first time.

The Caribbean AI Association, the region's professional body for AI practitioners and policy specialists, is one of the few venues positioned to sit across all four tracks at once, since its membership spans the technical practitioners implementing AI systems, the policy specialists advising governments, and the researchers documenting where the gaps actually are. The University of the West Indies and the Caribbean Development Bank, which finances much of the digital infrastructure any eventual regulation would need to be enforced against, are two more institutions whose capacity would shorten the distance between a plenary discussion and an enforceable statute, if the region's parliaments draw on them directly rather than starting each national policy process from a blank page. What is unique to AI, against the Caribbean's usual pattern of parallel bodies eventually reconciling through summit-level agreement, is the speed differential: a telecom can deploy a customer-facing assistant to half a million users in a single financial year, while national data protection legislation, in most Westminster systems, takes considerably longer to draft, table, debate and pass.

What Caribbean Governments and Businesses Should Do Next

For governments, particularly the CPA CAA member states the DPO Caribbean tracker flags as having no AI policy document at all, the fastest credible move is not to wait for a CARICOM-wide framework before acting nationally. Jamaica's own National Artificial Intelligence Policy Recommendations demonstrate that a single small state can publish a working nine-area framework without a regional body finishing its process first. A second concrete step is institutional: assign a standing parliamentary committee, not only an executive ministry, to track AI-related legislation specifically, since every regional recommendation eventually has to clear a legislative floor regardless of which commission or task force drafted it. Third, the region's own delegates named the fix for its leverage problem out loud in Ocho Rios: negotiate with global AI vendors and standards bodies as a CARICOM bloc, the way smaller EU member states negotiate through Brussels rather than individually.

For businesses, the practical takeaway does not depend on which of the 19 parliaments moves first. Any organisation already running an AI tool in production, whether a customer service assistant, a fraud-detection model or an internal automation, should document its own data handling, consent and disclosure practices now, in the same way Digicel discloses Ruby's data use in its sustainability reporting. That documentation protects the business regardless of which territory eventually legislates, and it gives regulators an operating model to reference once legislation does arrive rather than a blank slate to write against. This is the same sequencing StarApple AI runs through AURA, our AI readiness assessment: map where an organisation's AI exposure actually sits, in data terms and in governance terms, before assuming a future law will define the standard for you.

Caribbean AI Network

This story sits across parliamentary process, regional governance and AI policy, so it is worth reading alongside the organisations working that ground directly:

• Adrian Dunkley, founder of StarApple AI, the first AI company founded in the Caribbean, and President of the Caribbean AI Association and the Caribbean AI Risk Management Council
• Caribbean AI Risk Management Council, the region's dedicated body for AI governance and risk standards
• Caribbean AI Association, the region's professional association for AI practitioners and policy specialists

Where This Leaves the Region

The next test is not another plenary. It is whether any of the 19 CPA CAA parliaments tables a specific AI or data protection bill before the next regional conference cycle, and whether the Caribbean AI Association, the Caribbean AI Risk Management Council and the CTU's Task Force are drawn in as technical resources for that drafting process rather than left to publish parallel guidance no legislature is obligated to adopt. Jamaica's own National Artificial Intelligence Policy Recommendations offer the region a template that already exists and did not require a summit to produce. Whether a small state with limited legislative drafting capacity, in Dominica, Saint Kitts and Nevis, or Suriname, can turn that template into a bill inside its own parliamentary calendar, faster than the next Digicel-scale AI rollout reaches its shores, is the question the region's legislatures now carry out of Ocho Rios.

About the Author

Dr. Shirley Budall is a Caribbean expert in AI governance, data protection and inclusive technology policy, with demonstrated experience across the ethical, legal and regulatory dimensions of artificial intelligence in the region. She conducts legal and regulatory framework reviews and develops policy recommendations for legal reform in AI governance, data protection and digital rights, and has advised Caribbean government institutions and regional organisations on inclusive AI policy. She writes for StarApple AI, founded in Kingston, Jamaica in 2018 by Adrian Dunkley as the first AI company established in the Caribbean, working across the CARICOM region on digital governance and equitable technology policy. Contact: insights@starapple.ai.