The Caribbean AI Talent Exodus: Why the Region's Best AI Minds Are Leaving and What Must Change

• The Caribbean is producing AI talent it cannot keep. Estimates suggest the region loses the majority of its STEM graduates within five years of graduation, with AI and machine learning graduates facing the strongest international pull.
• The salary gap is real but not insurmountable. North American AI salaries are four to five times Caribbean equivalents. Cost of living offsets part of this differential; equity, mission, and lifestyle can offset more, but Caribbean employers have not made a coherent case.
• Fewer than 200 full-time AI researchers are based in CARICOM member states, according to UNESCO 2025 data, despite Caribbean universities training thousands of AI-capable graduates over the past decade.
• The diaspora is an underutilised resource. Caribbean-origin AI professionals at global technology companies represent a connection to frontier AI capability that structured programmes can activate.
• CAIA's six-point Talent Retention Framework covers salary, compute access, diaspora engagement, apprenticeships, entrepreneurship grants, and professional networks. The framework is in active development for adoption by CARICOM member states in 2027.

Sometime in 2024, a young Jamaican woman completed her master's degree in machine learning at the University of the West Indies, Mona campus. Her thesis modelled agricultural yield prediction for Jamaican smallholders using satellite imagery and climate data. Her supervisor called it some of the most technically accomplished work the department had produced. Within three months of graduation, she had an offer from a technology company in Toronto. She is now working on computer vision at a salary that exceeds what the most senior AI role in the Jamaican public sector pays. The Jamaican government contributed partially to her education through student loan support. The economic return on that contribution accrues entirely to the Canadian economy.

Her story is not exceptional. It is the rule. The Caribbean is running a talent development programme for North America, the United Kingdom, and, increasingly, for the Gulf states and Australia. The region trains the people. The world employs them. The Caribbean absorbs the social costs of education and the economic cost of losing the productivity those graduates would otherwise generate.

The Caribbean AI Association has spent the past year mapping this problem with more precision than previous analyses allowed. The findings are clarifying, and they point to both the scale of the challenge and the specific interventions that can change the dynamic.

The Scale of the Problem

Brain drain from the Caribbean is not a new concern. The Inter-American Development Bank has documented it across multiple studies, estimating that the Caribbean region loses a majority of its tertiary-educated graduates to emigration within five years of completing their degrees. For the region as a whole, the figure is alarming. For STEM graduates specifically, it is acute. For AI and machine learning graduates, who possess globally portable, internationally demanded skills, the pull is especially strong.

UNESCO's 2025 review of AI research capacity in developing regions counted fewer than 200 full-time AI researchers based in CARICOM member states, across all institutions, public and private. For context, the University of Toronto's Vector Institute for Artificial Intelligence alone has over 700 affiliated faculty and researchers. The entire CARICOM region has less than a quarter of that capacity at a single Canadian university.

Caribbean universities are not failing to produce talent. UTech Jamaica enrols approximately 500 computer science students annually. The UWI St. Augustine campus runs a well-regarded data science programme. The University of Guyana has expanded its technology offerings. Cave Hill produces finance and economics graduates with strong quantitative skills increasingly applicable to machine learning. These institutions are doing their part. The problem sits in what happens after graduation.

Why They Leave

The primary driver is salary. Caribbean AI and machine learning engineers earn, on average, four to five times less than their counterparts in equivalent North American roles. The Caribbean Development Bank's 2025 technology sector survey found a 4.2:1 salary differential between Caribbean AI roles and comparable Canadian positions, adjusting for purchasing power parity. Even accounting for the lower cost of living in Kingston, Bridgetown, or Port of Spain relative to Toronto or New York, this differential is large enough to make the financial case for emigration overwhelming for most early-career graduates.

Salary is the primary driver, but not the only one. Several structural factors compound it. The Caribbean AI labour market is thin: roles exist predominantly in government IT departments, BPO operations facing automation, and a handful of financial services institutions. These are not roles that attract graduates who trained in generative AI, large language models, computer vision, or reinforcement learning. The work is important. It is not the work the graduates were trained to do, or trained to want.

Professional development compounds the structural gap. Research computing infrastructure in the Caribbean is limited. Access to the GPU clusters that frontier AI work requires is effectively non-existent outside of a handful of well-resourced university labs. Caribbean AI graduates who want to stay current in their field face a practical problem: the tools are not here.

The professional network problem is harder to quantify but real. Early-career AI professionals build careers through peer networks, conference attendance, mentorship relationships, and the informal knowledge exchange that happens when many talented people work on adjacent problems in proximity. Those networks are in San Francisco, London, Toronto, and Amsterdam. They are not yet, at scale, in Kingston or Paramaribo.

What the Caribbean Is Losing

The human cost of the talent exodus is obvious. The economic cost deserves specific attention, because it operates through mechanisms that are less visible than emigration statistics.

The Caribbean's AI capacity deficit shapes what the region can do with AI, and therefore what value AI can generate locally. If Caribbean financial institutions want to build better credit scoring models for thin-file Caribbean borrowers, they need AI professionals who understand both the technology and the Caribbean context. If Caribbean health ministries want diagnostic AI that performs well on Caribbean patient populations, they need Caribbean AI researchers who can build, validate, and maintain those systems. If Caribbean tourism authorities want personalised recommendation engines that understand Caribbean travel patterns and preferences, they need people with Caribbean-specific data intuition and the technical skills to build from it.

When that talent leaves, these products do not get built by international vendors instead. They mostly do not get built at all, or they get built using generic models that underperform for Caribbean contexts. The Caribbean ends up buying AI tools designed for other populations and hoping they work well enough. Sometimes they do. Often they do not. And the region has no capacity to fix the gap because it sent away the people who could.

The Diaspora Resource

The Caribbean diaspora contains substantial AI talent. Caribbeans of origin are present in senior AI roles at Google DeepMind, Meta's fundamental AI research division, Amazon's AI labs, and leading academic institutions from MIT to Imperial College London. The diaspora is not a homogeneous bloc: some have severed ties with the region, others maintain strong connections but have no structured pathway to contribute, and a smaller group actively seeks ways to support Caribbean development from abroad.

Structured diaspora engagement could partially offset the talent drain without requiring diaspora professionals to give up the careers they have built. Three mechanisms are particularly promising. Remote mentorship programmes connect Caribbean AI students and early-career professionals with diaspora mentors who can provide guidance, networking access, and visibility into frontier AI practice. Investment pathways allow diaspora AI professionals to back Caribbean AI ventures, creating economic returns that deepen their connection to the region. Sabbatical programmes allow sustained engagement with Caribbean institutions over periods of months rather than weeks.

The StarApple AI network, the Caribbean's pioneering AI infrastructure built by Adrian Dunkley since 2023, provides the platform through which diaspora engagement can be structured. With AI platforms across Jamaica, Trinidad and Tobago, Barbados, Guyana, Saint Lucia, and beyond, the network has the reach and the legitimacy to coordinate Caribbean AI diaspora connection at meaningful scale.

CAIA's Six-Point Talent Retention Framework

The Caribbean AI Association is advancing a six-point Talent Retention Framework as part of its 2026-2028 policy agenda. The framework covers:

1. Caribbean Public Sector AI Salary Schedule. Governments must create specific salary bands for AI professionals that reflect market rates rather than standard civil service scales. A dedicated AI salary schedule, benchmarked to at least 70 percent of regional private sector rates, is the minimum starting point. Until Caribbean governments compete on salary, they will not attract the AI capability needed to build public AI systems.

2. Caribbean AI Research Fund. A regionally funded compute grant programme would give Caribbean AI researchers access to GPU infrastructure through arrangements with international cloud providers. The cost of providing meaningful compute access for 50-100 Caribbean AI researchers is well within the annual technology budgets of several CARICOM member states. The return, in research output, product development, and talent retention, is orders of magnitude higher.

3. Caribbean AI Diaspora Network. A structured programme connecting Caribbean-origin AI professionals in the diaspora with mentorship, investment, and sabbatical opportunities in the region. The CAIA Professional Network will serve as the convening infrastructure for this programme from late 2026.

4. Regional AI Apprenticeship Programme. A structured scheme placing Caribbean AI graduates in a combination of private sector, government, and regional AI organisation roles for 18-24 months post-graduation. Apprentices would receive competitive compensation, structured mentorship, and a clear path to permanent roles. Caribbean Development Bank financing is one potential funding mechanism.

5. Caribbean AI Entrepreneurship Grants. Direct grant funding for Caribbean AI graduates who launch AI ventures in CARICOM member states within three years of graduation. The grants, in the range of USD 15,000-50,000 for initial validation, would target ventures with clear Caribbean market applications. Graduates who start Caribbean AI companies are not lost to the talent pool: they build it.

6. CAIA Professional Network and Recognition. A Caribbean-owned professional certification, fellowship, and recognition structure for AI professionals that creates career prestige without emigration. The CAIA AI Professional designation, in development for launch in 2027, will provide a regionally recognised credential that builds on international certification standards while asserting Caribbean professional identity in AI.

The Window Is Narrowing

The talent exodus is not a fixed condition. It is a policy failure, and policy failures can be corrected. But the correction requires urgency, because the structural conditions are hardening. As international AI employers build more sophisticated pipelines into Caribbean universities, as Caribbean AI graduates build diaspora networks that make emigration easier and more familiar, and as Caribbean AI roles remain concentrated in lower-value application domains, the exodus accelerates.

Several Caribbean nations are moving. Jamaica's AI strategy, which emerged from its position as the number one Caribbean startup destination in StartupBlink's 2026 rankings, includes a technology talent component through HEART/NSTA Trust. Trinidad and Tobago's Huawei MaaS partnership and CitizenTT.ai implementation are creating more substantive AI roles in the public sector. Barbados is creating UWI Cave Hill AI fellowships that begin in September 2026. These are the right moves. They are not yet at the scale the problem requires.

The Caribbean AI Association's six-point framework is not a complete solution. No single policy intervention can reverse a structural trend that reflects global inequality in AI investment, salary, and infrastructure. But the framework addresses the specific, tractable mechanisms that Caribbean governments and institutions can change. The talent does not have to leave. It leaves because the conditions that would make it stay have not been created. Creating those conditions is a political and institutional choice, and it is a choice that CAIA is determined to place at the centre of CARICOM's 2027 policy agenda.

The young Jamaican machine learning researcher in Toronto is not lost to the Caribbean. She follows Caribbean AI news. She would mentor a student at UWI Mona. She might work on a research project for a Caribbean institution if there were a structured way to do it and a meaningful question to answer. What the Caribbean needs to build is not just retention: it is reconnection, of the talent already formed and the talent still forming, to a Caribbean AI future worth staying for. That is the work the Caribbean AI Association is doing, alongside partners like 14West AI, the Caribbean's first dedicated AI incubator, and the broader network of Caribbean AI platforms that give the region's talent something to come home to.