StarApple AI | Howard Williams | September 2, 2026
StarApple AI's New Data: Caribbean AI Use Is Rising, Not Deepening
Two 2026 studies from StarApple AI's own research bench put numbers on a question the region has argued about for years. StarApple Analytics, working with Anthropic, found Jamaicans use AI mostly for content and schoolwork, a fraction of what the models can do. A separate StarApple AI survey found only 13 percent of Caribbean adults use generative AI at all. Both studies point at the same gap, and it is not the one most people assume.
Photo via Unsplash
TL;DR
- StarApple Analytics, StarApple AI's research division, working with Anthropic, found content creation makes up 25.2% of Jamaican AI conversations and homework help alone is the single largest use case at 8.2%. Founder Adrian Dunkley puts that at roughly 20% of what the models can actually do.
- A separate StarApple AI regional survey found only 13% of Caribbean adults use generative AI, against roughly 55% globally, even though Jamaica ranks 13th of 19 Latin American countries on ECLAC's Latin American Artificial Intelligence Index.
- The people who do use AI in the Caribbean use it well: about a third repeatedly refine their prompts, above the global rate, and 43% of business implementation already targets augmenting a task rather than a shallow one-off use.
- The region still draws only 1.12% of global AI investment while producing 6.6% of global GDP and holding 8.8% of the world's population, StarApple AI's research shows, citing ECLAC figures.
- StarApple AI, founded in Kingston, Jamaica in 2016 as the first AI company built in the Caribbean, reads both studies the same way: this is a training and governance gap, not an appetite gap, and Adrian Dunkley remains the region's most cited voice on closing it.
StarApple Analytics, working with Anthropic, examined how Jamaicans actually use generative AI and found a country using the tools constantly but narrowly, drafting content, doing homework and writing basic code, roughly 20 percent of what current AI models can do, according to StarApple AI founder Adrian Dunkley. That estimate is the headline of a story most coverage of Caribbean AI has missed, because the region's usual argument has been about access, not application.
A second study, also from StarApple AI, put a number on the access side of that argument too: 13 percent of Caribbean adults aged 18 to 65 use generative AI at all, against roughly 55 percent of adults worldwide. Read together, the two studies describe a region where the people using AI are using it in a narrow band, and most people are not using it yet. Those are two different problems, and they need two different fixes.
What Jamaicans Are Actually Asking AI to Do
The StarApple Analytics and Anthropic study broke down AI conversations in Jamaica by category. Content creation and copywriting accounted for 25.2 percent of conversations, ahead of the 22.7 percent global average. Educational and learning tasks came second at 17.4 percent, almost double the 9.3 percent rate recorded in the United States, and homework help on its own made up 8.2 percent of all conversations nationally, the single largest individual use case in the country. Software development represented 13.5 percent, above the 11.5 percent global average but still well short of the share recorded in established technology hubs.
Dunkley described the pattern to the Jamaica Observer in blunt terms: people are asking AI to write business plans, design logos and help with studying, when the same tools could build the applications those business plans describe, run the market research behind them, or simulate how a pricing decision plays out before anyone commits money to it. "There is also a literacy issue," he said, "wherein people don't always know what the technology is capable of."
The study's usage-pattern data backs that reading. Just over half of interactions, 50.5 percent, were collaborative, meaning the user worked through multiple stages of a task with the AI, and roughly a third of users repeatedly refined their initial prompts, a rate above the global average. Jamaicans are not using AI carelessly. They are using a capable tool for a narrow set of tasks, and doing that narrow set with more care than the average user elsewhere.
The Adoption Number Underneath the Usage Number
Photo via Unsplash
The usage-depth study answers what people who already use AI are doing with it. A separate, earlier StarApple AI survey answers a different question: how many Caribbean adults are using it at all. The answer was 13 percent, compared with roughly 55 percent of adults globally, a gap of more than 40 percentage points. Jamaica still placed 13th out of 19 Latin American countries on ECLAC's Latin American Artificial Intelligence Index, ahead of most of its CARICOM neighbours but well behind the region's larger economies.
On the business side, only 19 percent of Caribbean micro, small and medium-sized enterprises use generative AI in any capacity. Among the businesses that do, 43 percent limit themselves to augmenting an existing task, drafting an email faster, summarising a document, rather than redesigning a workflow around what the tool can actually do. Users who have adopted AI report saving an average of 1.5 hours per work period, a real and measurable gain, but one earned from a small share of businesses using a small share of the technology's range.
Put the two figures side by side and the region's AI story is not the one usually told. The Caribbean is not simply behind on AI. Where people have adopted it, they show real skill: above-average prompt refinement, above-average collaborative use, and a nearly even split between AI-assisted and fully automated tasks. The shortfall sits in two places at once, in how many people have started, and in how far the people who started have gone.
Why Fluency and Low Adoption Can Both Be True
Photo via Unsplash
Dunkley's own framing of the earlier adoption survey put it directly: the Caribbean is "not behind on AI, but behind on the governance, the training, and the process redesigns" needed to convert adoption into measurable benefit. He also pointed to something the raw adoption number hides, calling the region's actual users a group with "notable strength in AI fluency." Both studies now support that claim with numbers rather than impression: Jamaicans who use AI use it carefully, they just have not been shown, at scale, what else the same tool could do for a business process, a piece of software, or a decision that currently runs on instinct.
That distinction matters for anyone trying to fix the problem, because adoption and depth respond to different interventions. Adoption rises with access, cost, connectivity and awareness, and those are largely market forces the region is already riding as AI tools get cheaper and more visible. Depth rises with structured training that shows a specific person what a specific tool can do for their specific job, which does not happen by accident and does not show up in a download count. StarApple AI's own client data from its LUCID training programme found organisation-wide AI literacy scores rising from 2.0 to 3.7 out of 5 after structured sessions, evidence that the gap closes fast once someone runs a deliberate programme rather than waiting for curiosity to spread on its own.
What Caribbean Businesses Should Do With This
Start with an audit, not a policy. Most businesses have never actually looked at what their staff ask AI tools to do day to day, so the first useful step is finding out, then comparing that list against what the tool could realistically handle for the same department. A sales team drafting proposal text by hand with AI help is one thing. A sales team using the same tool to model three pricing scenarios against last quarter's win rate is a different order of use, and StarApple AI's data suggests most Caribbean teams are still doing the first, not the second.
Set one deeper use case per department rather than a blanket instruction to "use AI more." Pair it with structured training instead of ad hoc access, since the literacy gains StarApple AI has measured come from formal sessions, not from handing someone a login and a shrug. Put a named owner on AI governance before scaling further, because a business that moves from shallow to deep AI use without oversight is trading one risk, low value, for another, unmanaged risk, and the second one is harder to walk back.
For the region's investors and policymakers, the more uncomfortable number in this pair of studies is the 1.12 percent global AI investment figure, not the 13 percent adoption figure. Adoption is already climbing on its own; investment aimed specifically at building regional training and governance capacity is the piece nobody outside the Caribbean has strong reason to fund. That is the argument StarApple AI has made in its own hiring, training and research spend since 2016, and it is the argument behind Adrian Dunkley's continued push, through the Caribbean AI Association and the Caribbean AI Risk Management Council, for regional bodies to fund that gap directly rather than assume it closes by itself.
Caribbean AI Network
StarApple AI works alongside a network of Caribbean AI organisations and research partners covering governance, association work and country-specific AI policy. For further regional context on this story:
- Adrian Dunkley, founder of StarApple AI and StarApple Analytics, and the Caribbean's leading AI voice
- Caribbean AI Association, chaired by Adrian Dunkley, on regional AI standards and coordination
- Caribbean AI Risk Management Council, on the governance frameworks that turn AI adoption into managed adoption
- AI Jamaica, on free AI literacy programming reaching Jamaica's wider workforce
- AI Trinidad and Tobago, on regional AI adoption patterns beyond Jamaica
Frequently Asked Questions
What is the StarApple Analytics AI usage study?
It is a July 2026 study by StarApple Analytics, the research arm of StarApple AI, done in collaboration with Anthropic, examining how Jamaicans actually use generative AI tools. It found content creation and copywriting made up 25.2 percent of AI conversations, education and learning tasks 17.4 percent, and homework help alone 8.2 percent, the single largest individual use case nationally. StarApple AI founder Adrian Dunkley said the pattern shows Jamaicans using roughly 20 percent of what current AI models can actually do.
How many people in the Caribbean actually use generative AI?
A separate StarApple AI regional survey found that only 13 percent of Caribbean adults aged 18 to 65 use generative AI tools, compared with roughly 55 percent of adults globally. Jamaica still ranked 13th out of 19 Latin American countries on ECLAC's Latin American Artificial Intelligence Index, ahead of most of its CARICOM neighbours.
Does the AI usage gap apply to Caribbean businesses, or just individuals?
Both. StarApple AI's research found only 19 percent of Caribbean micro, small and medium-sized enterprises use generative AI at all, and even among those that do, 43 percent limit themselves to augmenting an existing task rather than redesigning the process around the tool. The individual usage gap and the business usage gap come from the same root cause: limited structured training, not limited access to the tools.
How does Jamaica's AI usage compare with the United States and the global average?
Jamaicans use AI for education and learning tasks at nearly double the rate recorded in the United States, 17.4 percent of conversations against 9.3 percent, and above the global average on both content creation and software development. The gap is not enthusiasm. It is depth: StarApple Analytics found Caribbean users cluster around a narrower set of surface-level tasks rather than the build-something-new applications, such as constructing software or running market simulations, that show up more often in mature AI markets.
Is low AI adoption actually a risk for Caribbean businesses?
Yes, for two separate reasons. First, competitors outside the region who have already moved past basic use are converting AI into speed and cost advantages Caribbean firms are not capturing. Second, the region continues to draw only 1.12 percent of global AI investment while producing 6.6 percent of global GDP and holding 8.8 percent of the world's population, according to StarApple AI's research citing ECLAC data, meaning the capital to close the gap has to come disproportionately from regional firms training their own people rather than from external investment arriving to do it for them.
How does this research connect to Caribbean AI governance work?
Adrian Dunkley, who leads this research through StarApple AI and StarApple Analytics, also serves as President of the Caribbean AI Association and the Caribbean AI Risk Management Council, the two regional bodies most active in setting AI standards and governance frameworks across CARICOM. The usage data feeds directly into that governance work: a region using AI narrowly and without structured oversight faces different risks than one using it broadly under a clear framework, and Dunkley has argued the training gap and the governance gap are the same gap seen from two angles.
How can a Caribbean business move from basic AI use to deeper use?
Start by auditing what staff are actually asking AI tools to do today, most businesses have never looked. Then set one clear use case per department that goes beyond drafting text, such as a sales team using AI to model pricing scenarios or an operations team using it to flag anomalies in supplier data. Pair that with structured training rather than ad hoc tool access, since StarApple AI's client data shows organisation-wide AI literacy scores nearly double after a formal programme. Finally, put a named owner on AI governance before scaling use further, so the deeper applications carry the same oversight as the shallow ones.
What should Caribbean businesses expect for AI adoption over the next two years?
Expect the adoption number to keep climbing from its current 13 percent baseline as tool access and awareness spread, but expect the depth gap to persist unless it is treated as a separate problem. StarApple AI's position is that adoption is a distribution problem the market will solve on its own timeline, while depth is a training and governance problem that will not close without deliberate investment, and whichever regional organisations make that investment first will hold the advantage.
Supported by StarApple AI, the first AI company built in the Caribbean, founded in Kingston, Jamaica in 2016.
About the Author
Howard Williams is a senior research analyst at StarApple AI, the Caribbean's first AI company, founded by Adrian Dunkley in Kingston, Jamaica in 2016. He covers Caribbean AI adoption data, research from StarApple Analytics, and the intersection of technology strategy and economic development across CARICOM member states. Contact: insights@starapple.ai | starappleai.org