The Caribbean Jobs AI Will Reach First, and the Ones It Cannot Touch
The headline says six in ten Caribbean jobs are safe from AI. The headline is hiding the real story. The jobs most exposed are the formal, urban, office-based roles the region spent fifty years building toward.
In the Caribbean, the jobs most exposed to artificial intelligence are not the ones most people fear. The World Bank found in April 2025 that 30 to 40 per cent of jobs across Latin America and the Caribbean carry some exposure to generative AI, but only 2 to 5 per cent face direct automation risk. The exposed roles are mostly urban, formal, higher-paid, and office-based: clerks, call-centre agents, junior accountants, paralegals, and back-office staff.
The number that fooled everyone
When the IMF studied AI exposure in Latin America and the Caribbean, the first reading looked reassuring. Roughly six out of ten jobs appeared insulated from automation, according to analysis the UN Development Programme drew from IMF data in 2024. Ministers could exhale. Most regional work is manual, service-based, or informal, and machines that write emails do not pick mangoes, change hotel linen, or fix a bus.
That comfort is misplaced. The same body of research shows the problem is not how many jobs are exposed, but which ones. The UNDP analysis found that only about one in eight Caribbean and Latin American jobs is positioned to be made more productive by AI, against one in four in advanced economies. The region is less likely to be replaced wholesale, and also less likely to benefit. That is the worst of both outcomes: limited upside, concentrated downside.
The exposed jobs sit in a specific place. The World Bank's 2025 brief found that AI-exposed roles in the region are more likely to be urban, formal, higher-paying, and to require more education. These are precisely the jobs a young Jamaican, Trinidadian, or Bajan graduate spends years training for. The aspiration class is the exposed class.
How the risk actually works
AI does not erase a job in one move. It removes tasks. A job is a bundle of tasks, as the economist David Autor put it, and the ILO built its global index on exactly that idea. When enough of a role's tasks become automatable, the role thins out: fewer people do more, or the work is restructured into something else. Understanding the mechanism matters more than memorising a percentage.
Task automatability
How much of the role is routine text, data entry, calculation, or document handling. The ILO's 2025 index found clerical work remains the most exposed category worldwide.
Complementarity
Whether AI makes the worker faster rather than redundant. The IMF found only about one in eight regional jobs has high complementarity, against one in four in advanced economies.
Regional friction
Weak connectivity, informal work, and thin digital records slow both the threat and the benefit. The World Bank estimates roughly 17 million regional workers cannot access AI's productivity gains.
Human anchor
Physical presence, trust, care, and local judgement resist automation. A plumber, a nurse, and a hotel supervisor each hold tasks no model can perform from a server.
A tool arrives
A bank, BPO, or ministry adopts an AI assistant for drafting, data entry, or customer replies.
Tasks get absorbed
Routine writing, summarising, and lookups shift to the model. The human still checks the output.
Roles get restructured
One supervisor now reviews the work of three. Hiring slows before anyone is fired.
The job changes or thins
The role either becomes a higher-skill oversight job or disappears from the next budget. Outcome depends on whether the worker was retrained.
Caribbean AI analysis, applying the ILO task-bundle framework (ILO Working Paper 140, 2025) to regional conditions.
Why the Caribbean is a special case
Global studies are built mostly on rich-country data. The Caribbean breaks three of their assumptions, and each break changes the risk picture.
Informality blunts the threat and the benefit
Most Caribbean economies run on informal work. The ILO reported that over 91 per cent of workers in Haiti operate outside the formal economy, alongside 62 per cent in Barbados and 54.6 per cent in Jamaica. An informal vendor, fisher, or roadside mechanic is hard for AI to automate, because there is no digital record to feed a model. The same gap means these workers cannot easily use AI to grow either. Informality is a wall that blocks the storm and the sunshine equally.
Tourism concentrates the workforce in human-anchored roles
The ILO noted in 2025 that across the Caribbean, one in three tourism jobs is low-paid and seasonal. Housekeeping, food service, and grounds work are physically anchored and relatively safe from automation in the short term. The exposure inside tourism sits higher up: reservations, revenue management, marketing copy, and guest-service email. The front desk is more exposed than the room.
The gender split is sharper than the average suggests
The World Bank found women in the region are, on average, twice as likely as men to be in jobs at risk of automation. The ILO's global index shows why: clerical, financial, and customer-service roles carry the highest exposure, and women are over-represented in them. A regional average of 2 to 5 per cent hides a much higher number for women in formal office work in Bridgetown, Port of Spain, or Nassau.
The region's low average exposure is not a shield. It is a mask. Informality keeps the headline number down while leaving millions unable to benefit, and the exposed minority sits exactly where the region's formal, educated, female-heavy office workforce is concentrated. A low average can still mean a hard hit for the people the economy depends on most.
The Caribbean AI Job Risk Index
To make this concrete, we built a Caribbean AI Job Risk Index. It scores occupation clusters from 0 to 100, where a higher score means greater near-term pressure on the role. This is Caribbean AI analysis, not official national data. It combines three published frameworks (the World Bank's 2025 LAC exposure bands, the IMF's complementarity scores, and the ILO's 2025 occupational gradients) and adjusts each cluster for Caribbean conditions: informality, infrastructure, and sector concentration.
The score blends three inputs in equal weight: task automatability (how routine the work is), low complementarity (how little AI helps rather than replaces), and a regional friction adjustment (which lowers near-term risk where informality and weak connectivity slow adoption). A score above 65 signals high pressure; 40 to 65 signals real change ahead; below 40 signals relative safety for now.
| Occupation cluster | Caribbean example | Risk score | Band |
|---|---|---|---|
| Data entry and back office | Records clerks in a ministry or bank | 88 | High |
| Call centre and BPO agents | Outsourcing seats in Montego Bay or Georgetown | 84 | High |
| Bookkeeping and junior accounting | SME accounts staff, payroll clerks | 79 | High |
| Paralegal and document review | Junior staff in law and conveyancing firms | 74 | High |
| Marketing and content production | Agency copywriters, social media staff | 71 | High |
| Reservations and revenue admin | Hotel booking and rate-management desks | 62 | Moderate |
| Retail and bank tellers | Branch and shop counter staff | 58 | Moderate |
| Junior software and QA | Entry-level coders, test engineers | 55 | Moderate |
| Teaching and lecturing | Schools and tertiary institutions | 44 | Moderate |
| Nursing and allied health | Hospitals and clinics region-wide | 33 | Lower |
| Skilled trades | Electricians, plumbers, mechanics | 26 | Lower |
| Tourism service and hospitality | Housekeeping, food and beverage, grounds | 24 | Lower |
| Agriculture and fisheries | Farmers, fishers, field labour | 19 | Lower |
The pattern is clear. The high-risk band is full of desk jobs that pay above the regional median and require qualifications. The lower-risk band is full of physical, present, hands-on work. AI is coming for the keyboard before it comes for the toolbox.
How the risk lands across the region
Exposure is not evenly spread. A country's risk profile depends on what its people actually do for work. The tables below map the whole Caribbean, every CARICOM member state, the associate members, and the territories outside CARICOM, by dominant employment profile and the specific job types most exposed to AI. The logic: economies heavy in formal office work and outsourcing face more concentrated exposure; economies heavy in tourism, agriculture, and informal work face slower, more uneven change.
CARICOM member states
| Country | Dominant employment profile | Specific job types most exposed | Near-term pressure |
|---|---|---|---|
| 🇯🇲 Jamaica | BPO, services, tourism, public sector | Call-centre agents, records clerks, junior accountants, social media staff | High |
| 🇹🇹 Trinidad & Tobago | Energy, finance, public sector | Back-office finance clerks, payroll staff, insurance processors, government records officers | High |
| 🇧🇧 Barbados | Tourism, financial services, global business | Offshore banking admin, fund administrators, paralegals, clerical staff | High |
| 🇬🇾 Guyana | Oil and gas, public sector, agriculture | Government clerks, accounts payable staff, data-entry officers | Moderate |
| 🇧🇸 Bahamas | Tourism, offshore finance | Reservations agents, bank tellers, fund admin, guest-service email staff | Moderate |
| 🇸🇷 Suriname | Mining, agriculture, public sector | Ministry administrators, bookkeepers, clerical officers | Moderate |
| 🇧🇿 Belize | Tourism, agriculture, services | Hotel booking clerks, small-business bookkeepers | Moderate |
| 🇦🇬 Antigua & Barbuda | Tourism, services | Reservations staff, guest-service admin, marketing assistants | Moderate |
| 🇱🇨 Saint Lucia | Tourism, agriculture | Hotel front-desk and booking staff, clerical workers | Lower |
| 🇬🇩 Grenada | Tourism, agriculture (spice) | Booking agents, public-sector clerks | Lower |
| 🇻🇨 St Vincent & the Grenadines | Agriculture, tourism | Government admin, small office clerical staff | Lower |
| 🇰🇳 St Kitts & Nevis | Tourism, citizenship-by-investment, services | CBI processing clerks, hotel admin, paralegals | Moderate |
| 🇩🇲 Dominica | Agriculture, eco-tourism, public sector | Ministry administrators, booking clerks | Lower |
| 🇭🇹 Haiti | Informal economy, agriculture | Small formal office and NGO admin segment | Lower |
CARICOM associate members
| Territory | Dominant employment profile | Specific job types most exposed | Near-term pressure |
|---|---|---|---|
| 🇹🇨 Turks & Caicos | Tourism, offshore finance | Resort reservations, company-formation clerks, bank admin | Moderate |
| 🇮🇴 British Virgin Islands | Offshore finance, tourism | Company registry clerks, fund administrators, paralegals | High |
| 🇦🇮 Anguilla | Tourism, offshore services | Hotel admin, financial-services clerks | Moderate |
| 🇲🇸 Montserrat | Public sector, services | Government administrative staff | Lower |
Non-CARICOM Caribbean territories
| Territory | Dominant employment profile | Specific job types most exposed | Near-term pressure |
|---|---|---|---|
| 🇩🇴 Dominican Republic | Tourism, manufacturing, BPO (services ~62% of GDP) | Call-centre agents, free-zone admin, hotel reservations, bookkeepers | High |
| 🇵🇷 Puerto Rico | Services, pharma manufacturing, public sector | Insurance processors, bank clerks, paralegals, medical billing staff | High |
| 🇨🇺 Cuba | State sector, tourism, agriculture | State administrative clerks, tourism booking staff | Moderate |
| 🇻🇮 US Virgin Islands | Tourism, public sector | Hospitality admin, government clerical staff | Moderate |
| 🇧🇲 Bermuda | Insurance, reinsurance, finance | Insurance support staff, fund administrators, accounting clerks | High |
| 🇰🇾 Cayman Islands | Offshore finance, fund services, tourism | Fund administrators, banking clerks, paralegals, reservations staff | High |
| 🇦🇼 Aruba | Tourism, services | Hotel reservations, marketing assistants, clerical staff | Moderate |
| 🇨🇼 Curaçao | Tourism, finance, logistics, refining | Financial admin, port documentation clerks, reservations staff | Moderate |
| 🇸🇽 Sint Maarten | Tourism, cruise, services | Cruise and resort booking staff, hospitality admin | Moderate |
| 🇧🇶 Bonaire, St Eustatius & Saba | Tourism, public administration | Public-sector clerks, dive-tourism booking staff | Lower |
| 🇬🇵 Guadeloupe | Services, tourism, agriculture | Public administration clerks, banking and insurance staff | Moderate |
| 🇲🇶 Martinique | Services, tourism, agriculture | Administrative clerks, banking and insurance staff | Moderate |
| 🇲🇫 Saint Martin | Tourism, services | Hospitality booking and admin staff | Lower |
| 🇧🇱 Saint Barthélemy | Luxury tourism, services | Concierge admin, reservations and booking staff | Lower |
Three patterns stand out. The offshore-finance centres (British Virgin Islands, Cayman, Bermuda) carry high pressure despite small populations, because their economies run on exactly the document-heavy financial work AI handles well. The tourism-led islands cluster in the moderate and lower bands, with the exposure sitting in the booking and admin layer rather than the frontline. The largest Spanish-speaking economies, the Dominican Republic and Puerto Rico, combine big call-centre and services bases with high formal employment, which places them firmly in the high band.
Lower near-term pressure is not good news. It often reflects deep informality and weak digital infrastructure, the same conditions that lock workers out of AI's benefits. Haiti's low score is a sign of exclusion, not safety. The territories with the most to gain from AI are also the ones with the most to lose if they cannot get online and get trained.
What the global parallel teaches the region
The international studies are not just background. They are a preview, and the Caribbean can read the outcomes before they arrive.
The ILO's Working Paper 140 found that one in four workers worldwide is in a job with some generative AI exposure, but only 3.3 per cent fall in the highest category. Crucially, exposure is now rising in digitised professional roles, including software developers, data analysts, and financial analysts, not only clerks. The lesson for the Caribbean: even the "safe" graduate jobs are moving.
The World Bank found 8 to 12 per cent of regional jobs could gain productivity from AI, but up to half of those gains are blocked by missing digital infrastructure. Roughly 17 million workers cannot reach the upside. For the Caribbean, this is the central policy fact: the benefit is real but gated behind connectivity and skills the region has not yet built.
The IMF concluded that AI could help lift the region's long-stagnant productivity, but warned the region risks falling further behind advanced economies in AI adoption. Since 1980, regional incomes have not converged with the US. AI is the next test of whether that pattern holds or finally breaks.
Caribbean AI analysis, drawing on World Bank (2025), IMF (2024), and ILO (2025) exposure findings.
What workers, businesses, and governments should do
The honest message is not "AI will take your job." It is "AI will take some of your tasks, and the people who add the rest of the value will keep their jobs and earn more." The response splits cleanly by who is acting.
| Action | Who it protects and how | Difficulty |
|---|---|---|
| Learn to direct AI tools | A clerk who can prompt, check, and edit AI output becomes the supervisor, not the casualty. The skill is judgement, not typing. | Easy |
| Move toward human-anchored tasks | Client trust, negotiation, care, and physical work hold value. Workers should shift time toward what a model cannot do remotely. | Easy |
| Retrain the exposed BPO and clerical base | Employers can convert call-centre and back-office staff into AI-assisted roles before cutting headcount. Cheaper than rehiring later. | Medium |
| Close the connectivity gap | Governments that fix rural and small-island broadband let the 17 million excluded workers reach AI's upside. No connectivity, no benefit. | Medium |
| Build national AI skills programmes | Ministries of education and labour can embed AI literacy in schools, TVET, and universities so the next cohort enters ready. | Advanced |
| Protect women in clerical work | Targeted reskilling for the female office workforce, the group the World Bank flagged as twice as exposed, prevents a widening gender gap. | Advanced |
The decisive variable is not whether AI arrives. It is whether the region builds the connectivity and skills to turn exposure into productivity. The World Bank's finding that half the potential gains are blocked by infrastructure is, in policy terms, the whole game. Fix the pipes and the training, or watch the benefit flow to economies that did.
How well do you read the Caribbean AI jobs picture?
Five questions. Every answer is sourced.
Frequently asked questions
Sources
- World Bank, "Quantifying the Jobs Potential of AI in Latin America and the Caribbean," Results Brief, 15 April 2025.
- International Labour Organization, "Generative AI and Jobs: A Refined Global Index of Occupational Exposure," ILO Working Paper 140, May 2025.
- IMF, "What Can Artificial Intelligence Do for Stagnant Productivity in Latin America and the Caribbean?" IMF Working Paper 2024/219, October 2024.
- UNDP, "Riding the Digital Wave: Will Latin America and the Caribbean take its shot at reshaping productivity?" 2024, citing IMF (2024) exposure and complementarity data.
- ILO, "Beyond tourism: A policy framework for economic diversification and job creation in the Caribbean," May 2025.
- ILO 2026 Employment and Social Trends data, as reported by the Jamaica Gleaner, "The Caribbean labour market paradox," February 2026 (informality rates: Haiti 91%, Barbados 62%, Jamaica 54.6%).
AI will not take most Caribbean jobs. It will quietly take the tasks inside them, and reward the workers who learn to do the rest. The region's danger is not a wave of layoffs. It is standing still while the productivity gain flows to everyone who moved first. Adrian Dunkley, guest contributor · Caribbean AI Newsletter, May 2026
About the Author
This is a guest post by Adrian Dunkley, an artificial intelligence specialist focused on how the technology reshapes work, business, and productivity across the Caribbean. He writes and advises on applied AI for the region.
Read more of his work at adriandunkley.net.
About Caribbean AI
Caribbean AI is the official directory of artificial intelligence companies, labs, and innovators in the Caribbean. We exist to connect startups, enterprises, and researchers driving the region's AI growth.
We track who is building AI in the Caribbean, what is being built, and what the region needs to compete. From Jamaica to Curaçao, from Guyana to the Bahamas, we document the people, organisations, and ideas shaping the Caribbean's AI future.
