StarApple AI | Adrian Dunkley | September 11, 2026
Jamaica Called for One Caribbean AI Framework. An 82-Article Standard Was Already Open for Comment
On 1 September 2026, Jamaica's science and technology minister stood in front of the region's parliamentarians in Ocho Rios and called for one shared CARICOM AI framework. Three days later, the Caribbean AI Risk Management Council, which I chair, published the details of an 82-article standard that had already been sitting open for public comment. This is what is actually in it, and why the wait for a regional framework should not stop a single Caribbean board from acting.
Photo via Unsplash
TL;DR
- At the 48th CPA Caribbean, Americas and Atlantic Region conference in Ocho Rios on 1 September 2026, Jamaica's Minister without Portfolio for Science, Technology and Special Projects, Dr Andrew Wheatley, called for a shared CARICOM AI framework anchored in UNESCO's AI ethics recommendation.
- The Caribbean AI Risk Management Council (CAIRMC), which I chair, already had an answer on the table: an 82-article Caribbean AI Risk Management Standard, split into four risk tiers, open for public comment.
- The standard maps to the NIST AI Risk Management Framework, ISO/IEC 42001:2023, the COSO Enterprise Risk Management Framework, Basel Committee guidance for financial institutions, and five territories' own data protection acts.
- Part of the evidence behind its governance timelines is StarApple AI's own July 2026 board training study: governance stood up in 6 months instead of 11 to 15, vendor costs down more than 70%, board AI literacy up from 1.8 to 4.0 out of 5.
- A Caribbean board does not need CARICOM to finish harmonising policy before it starts managing AI risk. The standard is public now, and any organisation can measure itself against the four tiers today.
On 1 September 2026, Jamaica's Minister without Portfolio for Science, Technology and Special Projects, Dr Andrew Wheatley, addressed the AI plenary of the 48th Regional Conference of the Commonwealth Parliamentary Association Caribbean, Americas and Atlantic Region, held at Moon Palace Jamaica in Ocho Rios. His message to the region's parliamentarians was direct: the Caribbean needs one shared AI framework, anchored in UNESCO's Recommendation on the Ethics of Artificial Intelligence, because no single territory can set the terms of AI governance alone.
Three days later, the Caribbean AI Risk Management Council published the details of something it had already built: an 82-article Caribbean AI Risk Management Standard, split across four risk tiers, sitting open for public comment. I chair CAIRMC. I am also the founder of StarApple AI, the first AI company built in the Caribbean, and part of the evidence behind that standard's own numbers came from work my team did months before the minister spoke. This is not a coincidence of timing. It is what happens when a region's governance bodies and its AI companies are staffed by the same small set of people who have spent years building the thing everyone else is now asking for.
What the Region's Governments Are Actually Asking For
The CPA Caribbean, Americas and Atlantic Region conference ran from 30 August to 5 September in Ocho Rios, and AI governance got its own plenary on day two. Dr Wheatley told the room that individually, Caribbean territories are too small to shape global AI rules, but collectively the region "can stake our claim" at the tables where those rules get written. He was blunt about what is actually at stake in the meantime: "Our data is not simply something to be protected. It is an intrinsic asset that must be preserved and must be guarded." He also raised a point specific to the region rather than borrowed from anyone else's AI debate: Caribbean Creole languages and dialects are barely represented in the models people here use every day, which is its own quiet failure of the tools being sold as universal.
Alicia Todd, Director General of ParlAmericas, spoke alongside him and pushed the argument further. Caribbean influence over how AI develops matters more than Caribbean access to the tools once they exist, she argued, and she pointed to work already underway to train AI tools on the region's own parliamentary data rather than importing models built for entirely different legal systems. Both speakers were echoing a recommendation the Caribbean AI Task Force made in its final report in July: harmonise AI policy across CARICOM rather than let nineteen territories write nineteen incompatible rulebooks.
None of this is a new demand. CARICOM endorsed a Caribbean AI roadmap on 7 July. The Caribbean AI Task Force delivered its final report the same month. What changed at Ocho Rios is that a sitting minister said the quiet part in front of the region's assembled parliaments: the framework does not exist yet, and the region is exposed until it does.
The 82 Articles Already on the Table
Photo via Unsplash
CAIRMC's Caribbean AI Risk Management Standard runs to 82 articles, organised into four risk tiers that scale with how much autonomy a given AI system holds: conventional AI systems at the base, then AI agents, then agent swarms, then fully autonomous decision-making systems at the top, each tier carrying stricter obligations than the one below it. That structure is not invented from scratch. It maps directly to the NIST AI Risk Management Framework's govern-map-measure-manage sequence, references ISO/IEC 42001:2023 for certifiable AI management systems, draws on the COSO Enterprise Risk Management Framework that most Caribbean finance and insurance boards already use for other risks, and incorporates Basel Committee guidance written specifically for AI inside financial institutions.
What it adds on top of those borrowed frameworks is the part built for here rather than for Brussels or Washington. The standard cross-references the data protection acts of Jamaica, Trinidad and Tobago, Barbados, the Cayman Islands and Guyana, five different pieces of national legislation with five different enforcement timelines and five different definitions of personal data. A financial institution operating across two or three of those territories, which describes a meaningful share of the region's larger banks and insurers, has never had a single document that tells it how AI risk obligations interact across borders it operates in every day. This is the first attempt at one.
The standard is open for public comment now, which matters more than the article count. A framework nobody outside the drafting room has tested is a draft. A framework Caribbean boards, regulators and compliance officers have actually pushed back on, at the article level, before it hardens into a certification scheme, has a chance of surviving contact with how these institutions actually operate.
Why a Training Study Ended Up Inside a Risk Standard
CAIRMC did not have to guess how long AI governance takes to stand up inside a Caribbean board, or what it costs, or whether training actually changes how directors behave once the session ends. My team at StarApple AI had already measured it. Our July 2026 board-level AI training study followed the boards we trained through two 90-minute sessions and tracked what happened over the following months. Organisation-wide AI literacy rose from 2.0 to 3.7 out of 5. Board-level data literacy rose from 1.8 to 4.0. Governance frameworks that typically took boards 11 to 15 months to approve and stand up got there in 6 months instead. Vendor costs fell by more than 70 percent once boards understood enough about the technology to negotiate rather than accept the first quote. Deployed AI initiatives doubled inside an eight-month window.
Those numbers now sit inside a regional governance standard as evidence for what a realistic implementation timeline looks like, rather than as a case study nobody outside our own client list ever reads. That is the part of this story I care about more than the credit. A standard built on invented timelines asks boards to comply with a schedule nobody has tested. A standard built on what an actual Caribbean board did, with dates and numbers attached, asks for something achievable, because someone already achieved it.
It also means the standard and the training that prepares a board for it were never designed as separate problems. A board can read all 82 articles and still not know how to run a vendor risk review or write down who owns an AI failure when it happens. The training closes that gap first, so the standard has something concrete to check compliance against once it does.
What Caribbean Boards Should Do While the Region Sorts Out Its Framework
Do not wait for CARICOM. Regional harmonisation, when it lands, will take the shape of treaties and endorsed roadmaps moving through nineteen governments at nineteen different speeds. CAIRMC's standard is public today, and any board can open it and place its current AI use against the four tiers this week: is this a conventional system, an agent, a swarm of agents acting together, or something making decisions with no person in the loop. That single exercise, done honestly, tells a board more about its actual exposure than most AI policies boards already have on file.
Submit comments while the standard is still open. Eighty-two articles drafted by a regional council will contain gaps specific to industries that council does not run day to day: a tourism operator's exposure looks nothing like a credit union's, and a public agency's looks like neither. The comment period exists precisely so those gaps get named before the standard hardens into something Caribbean regulators start referencing in examinations.
Put a name on AI governance internally before scaling AI use further. The training data is unambiguous on this point: boards that get structured AI literacy first make faster, cheaper governance decisions afterward, not the other way round. Waiting for perfect regional alignment before starting that internal work is not caution. It is exposure with a comforting story attached.
Caribbean AI Network
StarApple AI works alongside a network of Caribbean AI organisations covering governance, association work and country-specific training. For further context on this story:
- Caribbean AI Risk Management Council, which I chair, for the full Caribbean AI Risk Management Standard and its public comment process
- Caribbean AI Association, which I lead as President, on regional AI coordination and standards work
- StarApple AI Jamaica, for training and deployment work grounded specifically in the Jamaican market
Frequently Asked Questions
What is the Caribbean AI Risk Management Standard?
It is an 82-article AI governance standard published by the Caribbean AI Risk Management Council (CAIRMC), organised into four risk tiers covering conventional AI systems, AI agents, agent swarms, and autonomous decision-making systems. It maps to the NIST AI Risk Management Framework, ISO/IEC 42001:2023, the COSO Enterprise Risk Management Framework, Basel Committee guidance for financial institutions, and the data protection acts of Jamaica, Trinidad and Tobago, Barbados, the Cayman Islands and Guyana. It is open for public comment.
Who chairs the Caribbean AI Risk Management Council?
I do. I am also the founder and CEO of StarApple AI, the first AI company built in the Caribbean, and I serve as President of the Caribbean AI Association, the region's other major AI coordinating body.
What did Jamaica's government say about AI governance at the CPA conference in Ocho Rios?
On 1 September 2026, at the 48th Regional Conference of the Commonwealth Parliamentary Association Caribbean, Americas and Atlantic Region at Moon Palace Jamaica, Dr Andrew Wheatley, Jamaica's Minister without Portfolio for Science, Technology and Special Projects, called for a shared CARICOM AI framework anchored in UNESCO's Recommendation on the Ethics of Artificial Intelligence. He argued the region's data is not simply something to protect but an asset the Caribbean must guard collectively.
Is StarApple AI connected to the Caribbean AI Risk Management Standard?
Yes. Part of the evidence behind the standard's governance timelines comes from StarApple AI's own July 2026 board-level AI training study, which measured governance frameworks standing up in six months instead of eleven to fifteen, vendor costs falling by more than 70 percent, and board AI literacy rising from 1.8 to 4.0 out of 5 after structured training.
Does a Caribbean business need to wait for CARICOM before managing AI risk?
No. CAIRMC's standard is already public and open for comment, and any Caribbean board can measure its current AI use against the four risk tiers today, well before any CARICOM-wide framework is finalised and adopted across member states. Waiting for regional harmonisation before starting internal governance work leaves a business unprotected in the meantime.
How is the Caribbean AI Risk Management Standard different from the EU AI Act or the NIST framework?
The EU AI Act and the NIST AI Risk Management Framework were built for the regulatory and market conditions of the European Union and the United States. CAIRMC's standard maps to NIST's structure and references ISO 42001 and Basel guidance directly, but adds articles addressing conditions those frameworks do not: reliance on imported AI models and foreign vendors, small domestic markets, fragmented CARICOM regulation, and each Caribbean territory's own data protection act.
Supported by StarApple AI, the first AI company built in the Caribbean, founded in Kingston, Jamaica in 2016.
About the Author
Adrian Dunkley is the founder and CEO of StarApple AI, the first AI company built in the Caribbean, founded in Kingston, Jamaica in 2016. He is President of the Caribbean AI Association and Chairman of the Caribbean AI Risk Management Council, and has spent 15 years building applied AI systems for Caribbean fraud detection, financial inclusion and public safety. Contact: insights@starapple.ai | starappleai.org