The Caribbean CoLabs AI resource hub
Learn what’s possible. Understand what matters. Find a useful next step for your world.
For the professional finding more time, the farmer reaching customers, the small business trying something new, and the organisation asking thoughtful questions.
Reviewed · Free to explore. No sign-up here.
01 / Learn at your pace
Start with one course, then practise on a low-stakes task. You don’t need a technical job to build useful AI skills.
Practise giving useful context, writing clearer prompts, reviewing answers and using AI responsibly. A practical introduction focused on ChatGPT.
Free course · 60–75 minutes
A ChatGPT account is required to enrol. Academy courses are available globally.
Explore what AI is, what it can do and where its limits lie. Learn through explanations and exercises without needing programming experience.
Free course · Self-paced
Sign up for full course access. Choose Introduction to AI as your starting point.
Follow AI literacy and practical technology learning pathways, with options to explore prompting, chatbots and responsible use.
Free learning · Choose your pathway
Registration is available through SkillsBuild. Check each course for its language and credential requirements.
Learn how AI creates content, how to protect sensitive information and how to check its output. Examples focus on Microsoft Copilot.
Free module · 6 units
Read without a profile; sign in to record progress and assessment results. Basic familiarity with Microsoft 365 is helpful.
Work through questions about fairness, human rights, transparency and accountability. Use exercises to connect the ideas to real decisions.
Free course · Self-paced
Public course readings are available online. See the course FAQ for exercise and completion requirements.
“Free” refers to the learning listed above, checked on 14 November 2026. Tool subscriptions, optional exams and certificates may have separate conditions. These are independent providers; inclusion does not imply a partnership or endorsement.
02 / Our region, in context
Understand what has been enacted, what is being proposed, and what is still a policy ambition.
AI-specific rules are only part of the picture. Existing privacy laws can matter when an AI tool handles personal information. The entries below are a selected, source-linked snapshot—not a complete register for every Caribbean jurisdiction.
Law or legal instrumentAn Act, decree or court direction. Scope and commencement matter.
Bill or announced planA legislative proposal or an intention to develop one. Neither proves enactment.
Policy or assessmentA strategy, roadmap or readiness exercise. Its publication does not make it a law.
6 places and regional groupings
The Act covers personal data and includes provisions on automated decisions, overseas transfers and data protection impact assessments. These provisions are relevant when assessing how an AI service will process people’s information; this is a general data protection Act, not a standalone AI Act.
Instrument: 2019 · Attorney General’s Office: Act (PDF) · See sections 18, 22–28 and 65.
The UN reported an evidence-gathering and validation workshop for a proposed Omnibus Digital Legislation Framework. Its October report described work before policy options and legislative design; it did not report enactment of an AI law.
Source: 15 October 2026 · UN Barbados and the Eastern Caribbean
A practical question: Where will your customers’ information go, and what safeguards apply when it leaves Barbados?
The Office of the Information Commissioner identifies 1 December 2023 as the Act’s implementation date. AI projects handling personal data need to be assessed against existing privacy obligations, alongside any future AI framework.
Source: 30 November 2023 · Office of the Information Commissioner
Practice Direction No. 1 of 2025 governs generative AI use in the courts it names, including the Supreme Court and Revenue Court. It addresses responsible, secure and transparent use in proceedings. This is a specific court instrument, not a general AI statute.
Effective date stated: 17 September 2025 · Supreme Court: practice direction (PDF)
The reconstituted AI Task Force was tasked with developing a draft policy and legislative framework. A June government update set November 2026 as the target for a draft policy. A target date is not evidence that the policy or a law has been adopted.
Sources: Office of the Prime Minister, 3 March 2026 · Jamaica Information Service, 4 June 2026
A practical question: Which rules already apply to your sector and the information you hold?
The Ministry of Public Administration and Artificial Intelligence is working with UNDP and UNESCO on AI landscape and readiness assessments. The work examines areas such as institutions, infrastructure, skills and legal frameworks to inform national policy. An assessment is not an enacted AI law.
Announcement: 16 January 2026 · Ministry release (PDF) · Current initiative page
A practical question: What skills, data quality and oversight would your organisation need before introducing AI?
The 2026 Speech from the Throne announced an intention to introduce an AI Governance Act, including a National AI Commission and ethical safeguards. The speech establishes the government’s plan; it does not establish that a bill has passed or a law is in force.
Source: 20 May 2026 · Office of the Prime Minister: Speech from the Throne
What to look for next: A published bill, parliamentary stages, the final instrument and its commencement provisions.
The government’s legal office reports that the decree of 2 November 2026 approves an AI ethics code for public services. It addresses transparency, data protection, human review and impact assessment for high-impact systems. Its stated scope is public institutions; it is not a general AI Act covering every business.
The official announcement gives institutions 180 days from publication to adapt their policies and procedures.
Instrument: 2 November 2026 · Executive Legal Counsel: official announcement (Spanish)
A Senate committee report of 26 May 2026 documents study of an ethical and safe AI use bill. Separately, a November Presidency announcement describes plans to develop a general public-sector AI regulation during 2027. These sources describe proposals and planned work, not proof of their enactment.
Senate committee report (Spanish) · Presidency: planned 2027 regulation (Spanish)
A practical question: Are you providing AI to a public institution, and what requirements apply to that particular use?
UNESCO reported that COTED-ICT members endorsed the Caribbean AI Policy Roadmap in 2026. It supports cooperation and national policy development, informed by regional consultations. Regional endorsement does not itself turn the roadmap into domestic law.
Source: 7 July 2026 · UNESCO / United Nations Caribbean
The CTU’s publications catalogue lists a Caribbean Artificial Intelligence Task Force Report dated 23 July 2026. Use the catalogue to explore the regional work alongside national sources; a task force report is not legislation.
A practical question: How can regional cooperation support Caribbean languages, skills, data stewardship and local problem-solving?
Reviewed 14 November 2026. Each entry states what its linked official source documents, with the source or instrument date. A later step may not yet be reflected here; an omitted country does not mean it has no rules. Check the current instrument, regulator and commencement details before relying on it. This is general information, not legal advice.
Suggest an update with an official source03 / Keep people in the picture
Who benefits? Who could be harmed? Who can question a result? These questions belong in everyday decisions.
An efficient tool can still give unfair answers, expose private information or make a service harder to access. Ethical use means considering those consequences before and after you introduce it.
Try local names, language varieties and different ways of expressing a request. Include the people who will use the service in testing it.
Ask what data is needed, where it goes, who can access it and when it is deleted. Use invented examples while learning.
AI can invent facts and references. Open the original sources, check dates and calculations, and have someone qualified review consequential output.
Decide who approves the result, who can stop the process and how someone can challenge a decision or get human help.
Make it clear when people are interacting with an AI assistant. Avoid presenting generated images, voices or testimonials as real evidence.
Consider accessibility, connectivity, cost and environmental impact. Keep another route available when an AI service would exclude someone.
Our practical prompts draw on UNESCO’s Recommendation on AI Ethics (2021, a voluntary international recommendation) and NIST’s AI Risk Management Framework and Generative AI Profile (2024 profile, voluntary guidance). The local examples are our suggestions for applying those ideas.
04 / Put learning into practice
These are illustrative exercises, not promises of results. Use an approved tool and keep sensitive details out.
Try a public customer message using invented stock details. Check prices, availability, collection arrangements and tone before publishing.
Help me draft a friendly WhatsApp update for a small produce business. Use this fictional example: tomatoes, BDS$6 per bag; cucumbers, BDS$4 each; collection Saturday, 9–11 am. Keep it under 80 words. Do not invent a location, health claim or delivery service. Flag anything I need to fill in.
Try a blank meeting template before using any real meeting notes. Review whether the structure fits your work.
Create a reusable agenda for a 30-minute project check-in. Include time for progress, obstacles, decisions and next steps. Leave names and project details blank. Add a simple action table with task, owner and due date. Keep space for questions and human discussion.
Use AI to prepare questions, then ask vendors for evidence. Compare the time saved with the time spent checking and correcting.
We are exploring an AI assistant that answers questions using our public website. Draft questions to ask a supplier about accuracy, sources, data storage, retention, model training, access controls, total costs and human handover. Include a small pilot plan and reasons to stop. Do not assume AI is the best solution.
05 / Keep this close
Use this as a conversation starter with yourself, your team or a technology partner. It is a practical planning checklist, not a compliance certification.
Download the checklist (.txt)Name the problem. What should become easier, and would a simpler tool do?
Set the data boundaries. What may be shared? Keep passwords, client records and sensitive information out of unapproved tools.
Check the service terms. Ask about storage, retention, training use, deletion, security and who can access the data.
Keep people involved. Who reviews output, catches errors and handles questions or complaints?
Test in your context. Check real language needs, accessibility, local knowledge and a range of possible users.
Count the whole cost. Include subscriptions, usage, setup, training, checking and ongoing support.
Try a small pilot. Use non-sensitive examples, define success and record what goes wrong.
Keep an exit route. Know how to pause, export your work, remove access and return to a workable alternative.
Caribbean CoLabs · caribbeancolabs.com/ai · Reviewed 14 November 2026
06 / A few things worth knowing
No. The courses linked above provide free learning. Start there, then decide whether a paid tool would solve a specific problem. Free learning does not mean every product demonstrated is free.
Start with invented information. For real records, first establish permission, applicable privacy and professional obligations, an approved service and appropriate access and data controls. Removing a name alone may not make information anonymous. Get qualified advice for sensitive or regulated work.
No. Clear instructions can help, but they do not eliminate errors or invented references. Check important claims against reliable originals. Asking an AI to check itself is not independent verification.
NIST’s discussion of confabulation (PDF)No. A strategy sets direction. A bill proposes legislation. An Act, decree or court direction has its own legal scope and effective date. Read the actual instrument and the stage reported by the official source.
That is a useful finding. A clearer process, a better template, connected tools or straightforward training may solve the problem more simply. Start with the outcome you need.
Learning is a good beginning
We help individuals, small businesses and organisations assess a useful starting point, work through data and cost questions, and build the skills to put a sensible plan into practice.
Talk through your AI questions connect@caribbeancolabs.com