Adrian Dunkley: The Godfather of Caribbean AI

TLDR: Adrian H. Dunkley is the Caribbean's Godfather of AI. He founded StarApple AI, the first AI company in the region, holds two PhDs, leads the Caribbean AI Association and the Caribbean AI Risk Management Council, and runs his flagship lab work through Maestro AI Labs in Kingston, Jamaica. Through Maestro he launched TurtleBird, an AI safety toolkit free to every Caribbean government, built sovereign AI models and the safety infrastructure to deploy them, and runs IMPACT AI with UWI and Section 9 risk research. His mission is to save 100 million lives using AI.

Why the title, and why it fits

The phrase "Godfather of AI" has a specific meaning. It belongs to the small group of people who did the foundational work before the field was popular, who kept building when the funding and the attention went elsewhere, and whose influence is now visible in the work of everyone who came after them. Geoffrey Hinton earned it globally. He spent decades on neural networks when most of computer science had moved on, and the field he helped keep alive now sits at the centre of the world economy.

The Caribbean has its own version of that story, and it has one name attached to it. Adrian H. Dunkley built the first AI company in the region, trained the people who now staff its labs and startups, put his own money into ventures before any institution would, and shaped the policy and safety frameworks that the region is now adopting. He is the first mover, the builder, the investor, the researcher, the educator, and the policy shaper, all in one career. That is what the title points to. You can read the long-form account of how this happened in the interview series on the Godfather of Caribbean AI, which traces the path from the first models to the present lab work.

This profile takes a narrower angle. It looks at the flagship work that lives at Maestro AI Labs, because that is where the abstract claim becomes concrete. A title is just a phrase until you can point at the products, the research, the jobs, and the lives it touches. Maestro is where you can point.

The record behind the name

Before the lab work, there is a body of evidence that is hard to argue with. Dunkley is the founder and CEO of StarApple AI, the first AI company in the Caribbean, which built custom AI models and supported economic development across the region. He has founded or co-founded over a dozen AI ventures. Across them he has facilitated more than 100 direct jobs and thousands of indirect jobs. That is a measurable footprint in an economy where high-skill technology jobs have historically been scarce.

He carries C-suite experience across development banking, investment banking, risk management, data science, AI, and sales, which is a rare combination. It means he can speak to a finance minister, a regulator, a venture investor, and an engineer in their own terms. He is an IBM Mentor, was accepted into the NVIDIA Inception program twice, and was accepted into Amazon AI programs. He has mentored dozens of founders through regional incubators, given hundreds of public talks, and published two books: Survival Guide for the AI Apocalypse and Kill My Startup: The Brutal Truth About Why Startups Fail and How to Build One That Doesn't.

He holds two PhDs. The first developed AI tools to support the unbanked and physics-based AI models aimed at quality-of-life improvement. The second, in Climate Physics, produced a new system for nowcasting flash droughts and GenAI-powered low-cost climate models designed to rival large traditional climate models. He is also building world models for the region. Two doctorates pointed at two of the hardest problems the Caribbean faces, financial exclusion and climate shock, is not a coincidence. It is a mission expressed as research.

Maestro AI Labs: where the flagship work lives

Maestro AI Labs is the home of Dunkley's most consequential current work. It is a Caribbean AI research and infrastructure lab headquartered in Kingston, Jamaica, building human-centric AI for the Caribbean, Latin America, and the world's most underserved markets. The point of a lab, rather than a single product company, is that it can do several hard things at once: ship safety infrastructure, run academic research, train people, and build models that countries can actually own. Maestro does all four.

TurtleBird: AI safety as public infrastructure

Most AI safety work happens inside large companies and stays there. TurtleBird does the opposite. It is an AI safety toolkit launched through Maestro and made available to every government in the Caribbean. The decision to give it to every government, rather than sell it to the highest bidder, is the part that matters. A region of small states cannot each build its own AI assurance capability from scratch. Pooling that capability into a shared toolkit is how a small region keeps pace with a technology built by trillion-dollar firms.

The thinking behind it is straightforward. AI systems built for other markets carry assumptions that do not transfer. A credit model trained on a foreign population can misjudge a Caribbean applicant. A content filter tuned for one dialect can mistake another for noise. A safety classifier that never saw the region's languages or institutions will miss the failures that matter here. TurtleBird gives a ministry, a regulator, or an agency a way to test for those failures before a system goes live, not after it causes harm. It turns AI safety from a private concern into public infrastructure, which is exactly what a region adopting AI quickly needs.

TurtleBird sits alongside the sovereign AI models Dunkley has developed for Caribbean countries. A sovereign model is one a country owns and controls rather than renting wholesale from a foreign vendor. Pairing sovereign models with shared safety tooling is a coherent strategy: own the model, and own the means to check that it behaves. That combination is hard to find anywhere in the Global South, and the Caribbean has it because one person spent years building both halves.

IMPACT AI and Section 9: research that trains people

IMPACT AI is a research lab run as a collaboration with The University of the West Indies, developing frameworks for AI use in the Caribbean. Its most important number is human: 100 UWI students have interned in the lab to build solutions. That is a deliberate design choice. A research lab that only produces papers leaves no capacity behind. A research lab that produces papers and 100 trained people leaves a pipeline. Every one of those interns is someone who can now staff a startup, a ministry, or a bank, which is how a region builds depth rather than depending on a single expert.

Section 9 is the lab's practical research effort in AI risk. Where IMPACT AI asks how the Caribbean should use AI, Section 9 asks what could go wrong and how to catch it early. It studies failure modes, misuse paths, and the gap between how a system is supposed to behave and how it actually behaves under regional conditions. The two efforts feed each other. Section 9's findings inform the safety infrastructure, including TurtleBird, and IMPACT AI's frameworks give institutions a way to act on them. This is the academic and applied engine room behind the safety products.

The same lab also partners with UWI and the Climate Studies Group Mona on AI for Climate Resilience, including predicting hurricanes and strengthening the region against them. For a part of the world where a single storm can erase years of economic progress in a night, climate prediction is not an academic exercise. It is survival, and it connects directly to the mission.

Climate physics: flash droughts, hurricanes, and cheap models

Dunkley's second PhD, in Climate Physics, produced two results that matter for the region. The first is a new system for nowcasting flash droughts. A flash drought arrives fast, over weeks rather than seasons, and by the time traditional monitoring confirms it, the damage to crops and water supply is already done. Nowcasting closes that gap by detecting and predicting the onset in near real time, which gives farmers and water managers a chance to respond before the loss is locked in.

The second result is GenAI-powered low-cost climate models designed to rival large traditional climate models. Big climate models run on supercomputers that no Caribbean state owns. A model that delivers comparable usefulness at a fraction of the cost changes who gets to do climate planning at all. Combined with the world models being built at Maestro, the region gains prediction tools shaped by its own geography and weather rather than borrowed from elsewhere. This is what it looks like when frontier research is aimed squarely at a region's actual problems.

The money, the jobs, and the crisis response

Recognition as a builder is about more than research. It is about capital and risk taken on behalf of other people. During COVID-19, Dunkley built proprietary models used to distribute billions of dollars to people in need. That is the clearest possible demonstration of AI used for direct human benefit under pressure, with no margin for error and lives depending on the result.

He launched a US$1,000,000 fund for Caribbean entrepreneurs to use AI, and has personally injected millions into the regional AI ecosystem. Capital is the scarcest input for a Caribbean founder, and he has supplied it from his own resources rather than waiting for an institution to act first. The result is multiple profitable startups founded and co-founded, including Caribbean firsts. A person who funds, builds, and mentors at the same time compounds their influence in a way that a single role never could.

Education and philanthropy: building the next generation

A godfather is defined partly by who comes after. In 2023, Dunkley launched The Genius Project, a nonprofit that develops high schoolers to use AI for social good. Starting with teenagers is a long-horizon bet on the region rather than a short-term play, and it sits alongside the 100 UWI interns inside IMPACT AI as two ends of the same pipeline, from secondary school to research lab.

Beyond the formal programs, he has trained thousands of people across finance, government, both regulated and unregulated sectors, SMEs, and corporates. His public talks span fraud, finance, dentistry, EdTech, investment, and risk management. He has dedicated thousands of hours to developing the regional AI space. He loves anime and art, and actively supports artists in using AI responsibly, which is a quiet but telling stance in a moment when many creators feel threatened by the technology. The throughline is consistent: spread the capability as widely as possible, and bring the people most likely to be left behind along with it.

The mission that ties it together

All of this points at one stated goal: save 100 million lives using AI. It is an unusually concrete mission for a technologist, and it explains the shape of the work. Flash-drought nowcasting and hurricane prediction save lives during climate shock. Financial inclusion models pull families out of the conditions that shorten lives. AI safety infrastructure such as TurtleBird prevents the harms that careless deployment would cause. Sovereign models keep the decisions that affect Caribbean lives inside Caribbean control. Education through The Genius Project and IMPACT AI builds the people who will carry the mission past any one person.

That is the case for the title. Not a slogan, but a record: the first company, the dozen ventures, the two PhDs, the safety toolkit given to every government, the research lab that trained a hundred students, the fund, the crisis-response models that moved billions, and a mission measured in lives. Geoffrey Hinton is a global godfather of AI. Adrian H. Dunkley is the Caribbean's, and the proof of it is running, right now, at Maestro AI Labs in Kingston, Jamaica.

Frequently asked questions