Adrian Dunkley | June 12, 2026
The Caribbean Can't Wait for Permission: The Case for Building AI Here
Most technology shifts of the past 60 years reached the Caribbean after they had already reshaped the economies that built them. With AI, the region can still take part in the building.
James Guy Evans, The Tow Boat Conqueror, 1852. The Metropolitan Museum of Art, public domain.
TLDR
- The assumptions built into AI systems now will shape how they work for the next 20 years. The Caribbean should help build them, not only buy the result.
- The region has real advantages: English-speaking talent in several language communities, time zones shared with North America, a large technical diaspora, and global reach in music, sport and the arts.
- The biggest obstacle is the habit of waiting for technology to arrive finished from elsewhere.
- StarApple AI, 14West AI, Maestro AI Labs and a network of 19 country AI platforms show the work has started.
- Concrete targets: 50 Caribbean AI companies above $1 million in revenue within five years, and a CARICOM AI framework within two.
The region has adopted every recent technology late
The internet reached most of the Caribbean in the mid-1990s, about five years after it began changing North American and European economies. Kenya had M-Pesa mobile money in 2007; Caribbean fintech firms were still fighting regulatory barriers to digital payments in 2020. Cloud computing became standard enterprise IT in the early 2010s, while many Caribbean organisations paid to run their own servers until the end of that decade.
Capability was never the cause. Distance from where the technology was made, weak investment in digital infrastructure, slow regulation and a habit of treating technology as an import all played a part. AI will show whether that habit has changed.
Three conditions that make AI different
Access is cheap. Anthropic's Claude API, OpenAI's GPT API, Google's Gemini and Meta's open-weight Llama models are available to anyone with a laptop and a payment card. A developer in Port of Spain uses the same models as one in San Francisco, which was never true of the infrastructure behind the internet, mobile or cloud eras.
The people exist. UWI, UTech and the University of Guyana turn out more computing and data graduates each year, and Caribbean professionals inside North American and European tech firms number in the tens of thousands. Remote work lets them contribute without leaving their jobs.
The rules are still being written. AI governance is being drafted at the ITU, UNESCO and the OECD and in national parliaments. If Caribbean governments take part, the standards can account for small island economies, tropical climate data and Creole languages. If they do not, rules written for Germany and California will be applied in Castries and Georgetown.
Four advantages the region undersells
- Language range. The region spans English, French, Spanish, Dutch and several Creoles. A Caribbean team building for Caribbean markets works across languages from day one, which teams in single-language markets pay heavily to learn.
- Cultural reach. Caribbean music, sport and art reach audiences far larger than the region. AI tools for music production, sports analytics or creative work built with that context can be sold well beyond the Caribbean.
- Local trust. Banks, hospitals and ministries are more willing to adopt AI from firms that know their data, their regulators and their accountability lines. A vendor 5,000 miles away that has never worked under a Caribbean regulator starts behind.
- A cluster of similar markets. The islands share enough law, culture and economics that a product built for one market can be adapted to several more at modest cost. That makes the region a test bed for small-economy use cases that exist all over the world.
What I have built, and the gap each piece fills
I founded StarApple AI in 2016 because the region had plenty of plans for technology leadership and no working example. The company had to show that a Caribbean AI firm could win clients, ship products and take part in global AI work from a Caribbean base.
14West AI exists because one company is not a sector. It is a fund and incubator that gives early Caribbean AI startups capital and mentoring until they are credible enough to raise outside money.
The network of 19 country AI platforms, from AI Jamaica to Saint Lucia AI to T&T AI, fills a simpler gap. There was almost no AI material written for Caribbean readers by Caribbean writers, and people cannot build companies in a field they have had no way to learn about.
Where the region stands, and targets worth setting
Market forecasts put global AI spending at around $826 billion by 2030. The Caribbean's share today rounds to zero. There are good signs: Jamaica ranked first in the Caribbean for startup activity in StartupBlink's 2026 index, its best result; Trinidad and Tobago deployed CitizenTT.ai; Barbados committed $187.8 million to a digital economy plan; and Guyana is building AI data centre capacity.
Those signs need targets. These are the ones I would hold the region to:
| Target | Deadline |
|---|---|
| A CARICOM AI governance framework with clear rules for business and government | 2 years |
| A Caribbean university producing 1,000 AI graduates a year | 3 years |
| 50 Caribbean AI companies with revenue above $1 million | 5 years |
| Caribbean representation in the senior leadership of a major global AI organisation | 5 years |
Each one depends on decisions Caribbean leaders can make now. The 1,000-graduate target is the hardest, because it needs lecturers the region does not yet have.
Stop waiting for permission
I have had the same conversation dozens of times with professionals who want to work in AI and are waiting: for better tools, more funding, a government programme, their employer to feel ready, or someone else to go first.
The tools improve every few months, so waiting for them never ends. Funding goes to people who have built something with early users. Government programmes tend to follow private activity. An organisation that waits until it is ready will find the target has moved.
I started in 2016 and have not regretted it.
What to do next
- Professionals interested in AI: build one working tool for a problem in your own job this month, using a paid API account (usually under US$20 to start), and show it to your manager.
- Procurement leads in banks, hospitals and ministries: require at least one Caribbean AI firm on every AI shortlist, and ask every vendor where your data is processed.
- Governments: name one official to represent the country in UNESCO, ITU and CARICOM AI governance meetings, and publish the country's position before the next regional summit.
- Universities: set a published target for AI graduates, and recruit diaspora practitioners as adjunct lecturers to close the teaching gap.
- Diaspora professionals: pick one Caribbean AI startup and give it a fixed commitment, such as a monthly advisory call or a small angel cheque.
Frequently Asked Questions
Can developers in the Caribbean sign up for the major AI APIs?
In most Caribbean countries, yes. Anthropic, OpenAI and Google each publish a list of supported countries, so check the list for your territory before you build. Billing is usually pay-as-you-go per token and needs an international card, which is often the practical barrier for young developers. Set a monthly spending cap in the account console on day one.
How can a small Caribbean government take part in global AI governance?
Start with UNESCO. Its Recommendation on the Ethics of AI was adopted by all member states in 2021, and it offers a Readiness Assessment Methodology that countries can request to map their laws and capacity. The results give a government a documented position to take into CARICOM and ITU discussions. Pooling representation through CARICOM is realistic for states without a dedicated AI office.
Does the Caribbean need to train its own AI models?
Rarely. Most useful Caribbean products can be built on existing models, with Caribbean documents supplied at query time and a Caribbean test set to check accuracy. Training or fine-tuning becomes worth it only when a firm holds a large, well-labelled local dataset, such as transcribed Patois speech. The test set is the asset to build first, because it shows where global models fail.
How should an organisation choose between a Caribbean AI firm and an international vendor?
Compare them on the same written checklist. Ask where data is stored and processed, who owns fine-tuned models and outputs, what support hours and response times are guaranteed in your time zone, and for a reference client under a Caribbean regulator. A local firm often wins on the last two, an international vendor on scale, and the contract terms decide more than the brand.
Is it too late for the Caribbean to catch up in AI?
No. Model access is the same price in Kingston as in London, so the gap is in applied products, data and skills rather than in the models. Those can be built over a few years with focused funding. What cannot be recovered is influence over rules that are finalised without Caribbean input.
About the Author
Adrian Dunkley is the founder and CEO of StarApple AI, the Caribbean's first AI company, and the founder of 14West AI, Maestro AI Labs, The Genius Project, and StarApple Analytics. He writes at adriandunkley.net.