96 Percent of Jamaican Students Already Use AI for Schoolwork. The Region's Rules Are Just Catching Up

Three things happened in Caribbean education in the space of nine days this July, and none of them were coordinated with each other. On July 7, the Jamaica Gleaner ran an editorial demanding a real national AI policy for schools. On July 13, the same paper carried a report on teachers nationally who are more worried about what AI is hiding from them than about students cheating with it. And on July 14, in Kingstown, Saint Vincent, the head of that country's National Centre of Technological Innovation confirmed his own government-backed AI pilot in schools, built on the opposite instinct: teach the tool, do not fence it off.

None of that is a coincidence, even if none of it was planned together. Two Ministries of Education, a university research team, a teachers' union, and a former principals' association leader all converged on the same subject in the same fortnight because the classroom had already made the decision for them. Research out of the University of Technology, Jamaica found that 96 percent of high school students already use AI tools for academic tasks. That policy conversation is playing catch-up to a number the classroom settled months ago.

The Number That Explains the Timing

Start with what UTech actually found, because it is the piece of data everything else in this story is reacting to. Professor Paul Golding, an information systems professor at UTech, and Tiou Clarke, a lecturer in the university's College of Business, surveyed students and faculty across Jamaica's education system on AI use. Their findings, reported in the Gleaner's July 7 editorial: 100 percent of university and college students surveyed use AI tools in their academic work, 96 percent of high school students do the same, and roughly 87 percent of faculty use AI programmes in class preparation and other professional tasks.

Read that last figure again, because it undercuts the easy version of this story, students racing ahead of institutions too slow to catch up. Faculty are close behind students, at 87 percent. The real gap sits between what is already happening inside Jamaican classrooms every day, at every level, and the fact that until this July there was no coherent written policy governing any of it. Professor Golding's own framing of the finding, quoted in the Gleaner, treats this as a curriculum problem, not a compliance one: "What this means is that the way we teach, what we teach, the way we assess how students learn, how we learn, will need to change."

What Jamaica Is Actually Piloting Right Now

The most concrete government response so far is a marking pilot. Jamaica's Ministry of Education, Skills, Youth and Information, under Minister Senator Dr Dana Morris Dixon, confirmed it is piloting AI-assisted marking of student assessments in a group of schools. The pitch is straightforward: automate the correction of routine assessments, generate real-time performance analytics that flag learning gaps earlier, and free teachers to spend the time they get back on lesson planning and direct student coaching instead of sitting over a stack of papers with a red pen.

That pilot does not sit alone. It runs alongside plans for a Jamaica Learning Assistant programme meant to adapt lessons to individual students, a Prompt Engineering course delivered through the HEART/NSTA Trust, and the country's first dedicated AI lab. Put together, the pattern is a Ministry treating AI less as a single tool to approve or ban and more as infrastructure to build across the whole system at once, assessment, tutoring, teacher training, and skills courses together.

The scale of the ambition is also why the policy gap the Gleaner flagged matters. A pilot in a handful of schools is a controlled experiment. A national rollout of AI-assisted marking, adaptive tutoring, and prompt engineering training, without a written policy governing data use, academic integrity standards, or age-appropriate access, is a much bigger bet to make on the fly.

Why the Teachers' Union Is Not Fully On Board

Mark Malabver, president of the Jamaica Teachers' Association, and Linvern Wright, a former president of the Jamaica Association of Principals of Secondary Schools, are not arguing against AI in education broadly. Both, according to the Gleaner's reporting, want limits specifically on how freely AI and connected devices reach younger children, at the early childhood and primary level in particular. Their reference point is a real, ongoing shift elsewhere: Sweden and Denmark have both moved in the past two years to scale back smart device availability in early classrooms, leaning back toward reading, writing, and handwriting practice on paper.

That comparison cuts against the instinct some Caribbean commentary has toward Asian systems moving the other direction, introducing AI tools at the earliest education levels rather than pulling back. The honest position, and the one the Gleaner editorial itself lands on, is that neither extreme is settled science yet. What is settled is that Jamaica currently has neither a Nordic-style restriction nor a fully built-out access framework. It has a widespread, largely unregulated default, 96 percent of high schoolers already using the tools daily, with the policy conversation only now catching up to describe rules for something that is already normal practice.

Saint Vincent's Different Bet

A day after the Gleaner's cheating-and-learning story ran, Petrus Gumbs, CEO of Saint Vincent and the Grenadines' National Centre of Technological Innovation, confirmed his organisation is working directly with that country's Ministry of Education to pilot AI use in Vincentian schools. His framing is deliberately different from a marking pilot or an access restriction. The goal, in his words, is to get students using AI "productively" rather than "willy-nilly," treating fluency with the tool as something schools should teach on purpose rather than either banning or ignoring.

Gumbs pushed back directly on the fear that AI use will make young people less capable, using a comparison that lands harder than a policy paper: "Persons who say that AI will make our children dumb, the technology has evolved. The dance hall has not emptied, but the DJ has evolved beyond record playing." The point underneath the metaphor: a tool changing how a skill gets practised does not erase the skill itself. A selector who moved from vinyl to digital decks stayed a selector. By that logic, a student who learns to use AI well keeps learning too, provided someone actually teaches them what "well" looks like.

Two territories, one shared week, two different instincts: Jamaica building marking infrastructure and pressing toward a formal policy; Saint Vincent building curriculum around productive AI use from the pilot's first design. Neither approach is finished, and neither government has claimed it is. What both share is the same underlying acknowledgment: AI in schools stopped being a future question sometime before either Ministry got around to writing the policy for it.

The Kingston Classroom Already Adjusting

Some of the most useful evidence in this story is not coming from a survey at all. The Gleaner's July 5 report described one English language and literature teacher at a Kingston-based high school who scrapped take-home writing assignments entirely after realising how many of her students had come to rely on AI to produce them. Her replacement is blunt and low-tech: extensive in-class writing, done by hand, in front of her, with no device in reach. Her stated worry sits past academic dishonesty: students who never struggle through a first draft on their own may never build the muscle a first draft is supposed to build in the first place.

That same reporting found a genuinely mixed picture among students themselves, not the runaway enthusiasm the adoption numbers alone might suggest. Between 50 and 60 percent of secondary students surveyed said they believe AI use is improving their "mental acuity." That leaves a substantial share who are either unsure or think the opposite, inside the same generation that adopted the tools almost universally. One CSEC candidate put it to me plainly during a summer programme session: "Mi use it fi check mi work, no fi do it. If AI write it, is not really me who pass, is a machine wearing my name."

What a Real Caribbean School AI Policy Should Actually Cover

Reading across everything that surfaced this July, from Kingston to Kingstown, a workable policy has to do more than function as a ban list or a permission slip. It needs to answer questions none of the reporting so far has fully settled: what data an AI marking tool retains on a student's performance, and who can access it. What counts as appropriate AI assistance at each grade level, rather than one blanket rule from primary through sixth form. How a teacher is meant to assess genuine understanding when a written assignment alone no longer proves it reliably. And what training teachers themselves receive before an AI tool is placed in their classroom, since 87 percent of Jamaican faculty are already using these tools without that structure fully in place.

This is also squarely governance territory, not just an education ministry question. Data retention, bias in automated marking, and accountability when an AI system gets something wrong all sit closer to the kind of AI risk framework organisations like the Caribbean AI Risk Management Council work on daily than to a curriculum committee's usual remit. A school AI policy written without that lens risks solving the pedagogy question while leaving the risk question wide open.

What This Means If You Are Heading Into a New School Year This September

CSEC results land in mid-August, and the school year that follows is the one most directly shaped by everything above. If your school adopts AI-assisted marking, your feedback loop on assessments gets faster, which only helps if you actually use that faster feedback to close gaps rather than just check a grade. If your teachers shift toward more in-class, device-free writing, the way Golding's own research and that Kingston classroom both suggest is already happening, treat it as a skill-building opportunity rather than an inconvenience. The students who come out ahead in a year like this one will be the ones who can clearly explain, without AI, everything they claimed to know with it, not the ones who simply used it the most.

None of this policy conversation is finished, and nobody covering it this July claimed it was. What changed is that the gap between practice and policy, 96 percent adoption against zero written rules, finally got named out loud, in two countries, in the same nine days. The Genius Project exists for exactly this moment: teaching Caribbean teens to use AI well before the rules are fully written, so that whichever policy each Ministry lands on, the habits already built are the ones that hold up.