Caribbean Jobs Most at Risk From AI in 2026: By Sector, By Country, and What to Do About It

Caribbean AI  /  The Boardroom Brief  /  Workforce

AI exposure across the Caribbean is real, measurable, and unevenly distributed. BPO, junior clerical, and customer-service roles sit in the highest-exposure category. Tourism front-line and skilled trades sit at the bottom. The window for getting Caribbean workers on the augmentation side of that line is the next eighteen months.

23 June 2026 14 min read Caribbean AI Newsletter
1 in 4
Share of global workers in occupations with GenAI exposure, per the ILO-NASK 2025 Global Index
2 to 5%
Latin American and Caribbean jobs at risk of full automation, per the ILO-World Bank 2024 study
~62,000
Jamaicans employed in BPO, the largest single concentration of high-exposure work in the region

AI exposure in the Caribbean concentrates in three job categories: BPO and call-centre work, junior clerical and back-office roles, and entry-level customer service. About 1 in 4 jobs across the region falls in GenAI-exposed occupations; 2 to 5 percent face full automation risk. Tourism front-line, skilled trades, and clinical health roles sit outside the highest-exposure tier.

Executive summary
Situation

GenAI exposure now sits at 1 in 4 jobs globally, and between 26 and 38 percent across Latin America and the Caribbean depending on subsector. Caribbean BPO, the largest single concentration of high-exposure work, employs well over 100,000 people across Jamaica and the Dominican Republic alone.

Complication

The highest-exposed Caribbean jobs are disproportionately held by women, formal-sector urban workers, and recent graduates. Existing labour-market institutions cover none of these cohorts well. The window for designing transition support before political cost outpaces feasibility runs out before the next election cycle in most member states.

Resolution

Five moves, in this order: run a workforce exposure audit, stand up AI literacy for highest-exposure staff, fund a regional skills transition mechanism, retrofit tertiary curricula for AI augmentation, and publish a national AI labour transition policy this year.

Caribbean job exposure concentrates in three categories, not spread evenly

The first finding from the recent ILO and World Bank studies on the Caribbean is that GenAI exposure does not spread evenly across the regional labour market. Three categories carry most of the risk. Call-centre and BPO work sits at the top of the exposure index because the tasks involved (scripted customer interactions, routine query resolution, data lookup, ticket classification) map almost exactly onto what large language models do best. Junior clerical and back-office work (data entry, document processing, scheduling, report drafting) sits close behind, for the same reason: the work is structured, predictable, and largely text-based.

The third category is entry-level customer service across other industries, including financial services, telecommunications, and utilities. Exposure here is moderately high rather than top-tier, because the work mixes routine handling with judgement calls and relationship maintenance that current models do not yet handle reliably. The 2 to 5 percent figure for full-automation risk applies mainly to the top exposure tier; the rest of the regional workforce sits in lower-exposure categories where AI augments rather than replaces.

What this means for the Caribbean is that exposure is concentrated by sector and by country, rather than spread evenly across the economy. Jamaica's profile (BPO-heavy, services-anchored) looks nothing like Guyana's (energy-dominant, public-service-heavy), and Trinidad and Tobago's profile differs again from Saint Lucia's. The regional response has to be country-specific even when the underlying technology is uniform.

Exhibit 1
Caribbean job categories ranked by AI exposure level
Relative exposure of Caribbean job categories, mapped to the ILO four-tier exposure framework.
BPO and call centres Tier 1 Junior clerical and back-office Tier 1 Junior content and translation Tier 2 Entry-level customer service Tier 2 Junior IT support and junior software Tier 2 Bookkeeping and finance support Tier 2 Paralegal and legal research Tier 3 Junior journalism Tier 3 Teaching and training Tier 3 Tourism front-line and hospitality Tier 4 Healthcare clinical Tier 4 Skilled trades Tier 4 Low Moderate High AI EXPOSURE LEVEL

Source: Author's mapping of Caribbean job categories to the ILO-NASK 2025 Refined Global Index of Occupational Exposure to GenAI. Tier 1 corresponds to the highest exposure category in the ILO framework; Tier 4 the lowest.

Where Caribbean intervention has to land first: the exposure-employment map

A two-by-two view of the region's risk profile clarifies which interventions justify the most political capital. The horizontal axis measures AI exposure (how automatable the work is). The vertical axis measures Caribbean employment share (how many regional workers do that work). The intersection at the top-right is the priority zone, where exposure is highest and employment is largest. Almost everything in that quadrant is BPO, junior clerical, and customer-service work.

Exhibit 2
The Caribbean priority zone: high exposure, high employment
Caribbean job categories positioned by AI exposure and regional employment share.
STABLE PILLAR PRIORITY ZONE NICHE PROFESSIONAL RISK NON-PRIORITY BPO and clerical ~100k+ workers regionally Customer service Tourism front-line Region's largest employer Skilled trades Junior IT & content Junior legal & finance Small specialty manual AI EXPOSURE Low High CARIBBEAN EMPLOYMENT SHARE Low High Top-right quadrant is where political capital should land first.

Source: Author's analysis based on ILO-NASK 2025 occupational exposure data and Caribbean labour-force composition data from national statistics offices and Caribbean Development Bank.

Three quadrants outside the priority zone matter for different reasons. The top-left is the region's stable pillar: tourism front-line, hospitality, manual services, skilled trades, and in-person care. These jobs face other economic risks, but AI is not the one to plan for. The bottom-right is niche professional risk: small cohorts of junior journalists, junior legal staff, junior software engineers, and junior accountants. Displacement per worker is real; aggregate regional impact is small. The bottom-left is not an AI-specific policy priority.

Country-by-country exposure profile across all sixteen Caribbean states

The exposure profile differs sharply across the sixteen Caribbean countries the directory covers. The table below maps each country's largest employment concentration to its highest-exposure category, with a recommended mitigation timeline.

CountryLargest employerHighest-exposure categoryWindow
JamaicaTourism & BPO (~62k BPO workers)BPO and call-centre workThis year
Dominican RepublicTourism, Manufacturing & BPO (~36-40k)BPO and back-office servicesThis year
Trinidad and TobagoEnergy and public sectorPublic-sector clerical and junior professionalQuarter
BahamasTourism and financial servicesBanking back-office and junior legalQuarter
BarbadosTourism, financial services, ICTJunior financial services and IT supportQuarter
GuyanaOil and public servicePublic-sector clerical (oil-sector roles low exposure)Watch
SurinameMining, services, public sectorPublic sector and small Dutch-market BPOQuarter
BelizeTourism, agriculture, small BPOCustomer service and agricultural adminWatch
Saint LuciaTourism and servicesHospitality back-office and junior professionalWatch
GrenadaTourism and agricultureClerical and customer service (small base)Watch
Saint Vincent and the GrenadinesTourism and agricultureClerical and hospitality back-officeWatch
Antigua and BarbudaTourism and financial servicesJunior finance and customer serviceQuarter
DominicaTourism and agricultureSmall BPO and clericalWatch
HaitiAgriculture and garment manufacturingPublic-service clerical (garment work low exposure)Watch
CuraçaoTourism, finance, refiningJunior finance and customer serviceQuarter
ArubaTourismHospitality back-office and customer serviceQuarter

Reading across the table, the countries with the largest absolute exposure are also the countries with the largest BPO and clerical concentrations: Jamaica and the Dominican Republic. The countries with the lightest exposure are the smaller agricultural and tourism-dependent OECS states, where the high-employment work is physical and in-person. Trinidad and Tobago, the Bahamas, Barbados, and Curaçao sit in the middle band, with concentrated exposure in their financial-services back-office layers rather than across the economy.

Four economic typologies determine the playbook

Across the sixteen countries, four economic profiles drive the AI exposure picture and the mitigation playbook that should follow.

Tourism-anchored small services economies (Antigua and Barbuda, Aruba, Barbados, Curaçao, Dominica, Grenada, Saint Lucia, Saint Vincent and the Grenadines, and the Bahamas) have lower top-tier exposure because tourism front-line work resists AI substitution. Their AI risk concentrates in back-office, marketing, finance, and customer-service operations that support tourism. The mitigation priority is upskilling the back-office layer, not retraining the front-line.

BPO-heavy services economies (Jamaica primarily, the Dominican Republic secondarily, with smaller BPO presence in Belize, Guyana, Suriname, and Trinidad and Tobago) face the largest absolute exposure. Tens of thousands of jobs sit in directly substitutable categories. Mitigation here is national-scale and time-bound to the next eighteen months. Jamaica's SAFE Task Force, Decent Work Recognition Programme, and Unemployment Insurance Benefit (Cabinet-approved May 2025) are the closest thing the region has to a working template.

Energy and commodity economies such as Trinidad and Tobago (gas), Guyana (oil), Suriname (mining and oil), and to a lesser extent the Dominican Republic (mining), have lower direct AI-displacement risk because the primary work is physical extraction. The risk concentrates in the public sector, the financial sector that supports it, and the professional services that orbit it.

Mixed-economy middle-income states, namely the Dominican Republic (manufacturing plus tourism plus BPO plus agriculture) and Haiti (agriculture plus garment manufacturing plus public service), face mixed exposure profiles where AI policy has to address several sectors at once. The BPO and clerical layer is the first priority in both, even though it is not the largest employer in either.

Skills before AI: what mattered first matters more

Most AI-readiness programmes start with the tools, hit a ceiling within six months, and then circle back to the deeper skills they should have led with. Reversing that order matters. The capabilities that mattered before AI now matter more, because they separate the worker who directs the model from the worker the model replaces. Five capabilities sit in this category, and any AI-readiness curriculum has to cover them before it covers prompting.

Skill 1

Domain expertise

Knowing what good work in your field looks like. AI fills the page; only domain expertise tells you whether the page is filled with sense or with confident nonsense.

Skill 2

Critical reading at speed

Verifying claims, spotting weak logic, catching what the model got subtly wrong. The most important skill for surviving the published critical-hallucination rates.

Skill 3

Plain written communication

Turning AI drafts into something a person would read. The model writes a paragraph; the worker turns it into the sentence the reader needed.

Skill 4

Judgement under uncertainty

Deciding when to use the model's answer, when to override it, when to escalate. Most professional decisions involve weighing competing constraints, and that weighing is the part the model cannot do for you.

Skill 5

Ethical reasoning

Knowing what should not be done with AI, even when it can be done. Professional liability still falls to the human in the loop, which is the consequence most worth taking seriously.

Skills beyond tools: what the next decade rewards

Once the foundation skills are in place, a second tier of capabilities determines who captures the value above the median. These are the skills that move a worker from the worker the model assists to the worker the model multiplies.

Skill 6

Instruction design

Getting the right output the first time. Prompting is a skill the way writing a brief is a skill, and the same people who write good briefs write good prompts.

Skill 7

Workflow architecture

Knowing where AI fits and where it should not. Teams in the StarApple AI study that integrated AI into their workflows captured roughly three times the value of teams that appended it.

Skill 8

Quality assurance

Catching hallucinations before they reach decisions. The cost is a few minutes per task; the avoided cost is materially larger and falls on the institution every time it gets skipped.

Skill 9

Cross-domain synthesis

Connecting ideas across fields, which the model handles poorly because its training data was built around existing disciplinary boundaries. The work AI does worst is also the work that pays best.

Skill 10

Relationship trust

What AI cannot fake. The Caribbean work that travels well is the work where reputation, accountability, and lived context matter, and those do not survive being automated away.

The Caribbean Godfather of AI on the moment

Every previous automation wave handed the Caribbean a clean replacement path. Manufacturing went, BPO came, BPO is going, and AI is the first cycle where the substitution and the augmentation are inside the same product. The worker who learns to direct it sees materially higher pay than the worker doing the same job by hand. The policy window for getting more Caribbean workers on the right side of that line is this year, not the next election cycle. Adrian Dunkley  ·  Founder, StarApple AI  ·  The Caribbean Godfather of AI

Three mitigations that work at scale

Three mitigation moves have evidence behind them at regional scale. None of them are speculative; all of them have early implementations either inside the Caribbean or in comparable economies.

Tier-1 workforce uplift through national AI literacy programmes
Mitigation 1

Direct AI literacy training for the highest-exposure cohort: BPO agents, junior clerical staff, and customer-service workers. Unit cost is a few hundred USD per worker; the alternative is a transition cost an order of magnitude larger, paid mostly by the public purse. Jamaica's SAFE Task Force (2025), Decent Work Recognition Programme, and the Cabinet-approved Unemployment Insurance Benefit are early-stage versions of this approach and the most replicable model in the region.

Education pipeline retrofit for AI augmentation
Mitigation 2

Tertiary curriculum updates for AI augmentation in business, accounting, law, journalism, and software programmes. Already underway at UWI and Anton de Kom University, but not yet at scale. The lag between curriculum change and labour-market impact is three to five years, which is why this has to start now rather than next year. The University of the West Indies' campus-wide AI literacy initiatives are a working template for other regional institutions.

Real-time labour-market signal infrastructure
Mitigation 3

Country-level monitoring of which job categories are losing hours and which are gaining. The data system that does not yet exist in any Caribbean state, and without it governments will not learn about labour-market disruption until the layoffs have already happened. The Caribbean Development Bank and CARICOM Secretariat are the natural anchors for a shared regional version, with national statistics offices providing the feed.

The five-step pattern governments and employers can follow

A workforce AI transition playbook for ministers, BPO operators, and university leadership
1
Measure
Run an exposure audit on your workforce covering roles, tasks, and timeline. Without the audit, every subsequent step is guesswork.
2
Reskill
Stand up AI literacy for tier-1 staff, paid for and structured by the employer, or by an industry consortium for smaller firms.
3
Augment
Wire AI into existing workflows with human review at the decision points. Augment the worker; do not run AI in parallel.
4
Reallocate
Move freed-up capacity to higher-value work the same workers can do with their existing experience. The augmentation case lives or dies here.
5
Govern
Publish a transition policy with timelines, metrics, and a defensible audit trail. The political insurance is worth the documentation cost.

Source: Author synthesis informed by ILO-NASK 2025 framework, IMF SDN/2026/001, and StarApple AI three-month Caribbean Claude study, 2026.

What Caribbean governments and employers should do this year

MoveReturnWindow
Conduct an AI exposure audit of the workforce, public sector first, then BPO and financial servicesIdentifies which roles need urgent intervention and provides the data for any subsequent policyThis month
Launch AI literacy programmes for highest-exposure staff (BPO, clerical, customer service)Two to three times the value per worker; meaningfully reduces displacement riskThis quarter
Fund a regional skills transition mechanism through CARICOM and Caribbean Development BankSmooths labour-market shocks before they hit; political insurance for ministers of labourThis year
Update tertiary curricula for AI augmentation across business, law, accounting, software, journalismPipeline correction; three to five year lag to labour-market impact, so this has to start nowThis year
Publish a national AI labour transition policy with timelines and metricsCoordination signal to industry, investor confidence, and Board-defensible audit trailThis year
Build real-time labour-market exposure monitoring at the national statistics officeGovernment learns of disruption before layoffs happen, not afterYear-plus
For Ministers of Labour and Education, BPO leaders, and university leadership

2026 is the policy year, not 2027

The honest question for Caribbean governments in 2026 is no longer whether AI will affect employment in the region. The ILO and World Bank measurement has settled that, and the IMF research has settled the timing. The open question is whether the Caribbean workforce gets prepared for the augmentation side of the curve or gets left to absorb the displacement side without institutional support. Most member states have not yet published a national AI labour-transition policy. Most member states should this year.

Reader test

How well do you know the Caribbean AI jobs picture?

Five sourced questions.

1. What share of jobs globally are in occupations with GenAI exposure, per the ILO-NASK 2025 Global Index?
2. What share of Latin American and Caribbean jobs are at risk of full automation, per the ILO-World Bank 2024 study?
3. Which Caribbean sector has the highest concentration of AI-exposed jobs?
4. Approximately how many Jamaicans work in the BPO sector?
5. According to the IMF SDN/2026/001 research, which cohort faces the highest AI displacement risk?
0/5

Frequently asked questions

BPO and call-centre work, junior clerical and back-office roles, and entry-level customer service across financial services, telecommunications, and utilities. The ILO-NASK 2025 framework places these in the highest exposure category. Tourism front-line, skilled trades, and most clinical health roles sit in lower-exposure tiers.
Lower than high-income countries on average (34 percent in high-income versus 25 percent globally in the ILO measurement) but with high concentration in specific Caribbean job categories. Latin America and Caribbean as a region falls between 26 and 38 percent depending on sub-sector, per the ILO-World Bank "Buffer or Bottleneck?" study, 2024.
Because clerical and customer-service occupations, which are the most exposed, have higher female employment share in most economies. The ILO measurement in high-income countries shows 9.6 percent of female employment in the highest-exposure category against 3.5 percent for men. Caribbean BPO and clerical work follow the same pattern.
Tourism front-line work (hospitality, food service, in-person guest interaction) sits among the lowest-exposure categories because the work requires presence and physical task completion. Tourism back-office (bookings, marketing, customer service, finance) is more exposed and should be in the mitigation plan, especially in tourism-dependent states.
The sector will not collapse, but it will transform. Caribbean BPO operators that move from voice-only customer service to AI-augmented service (where agents handle the complex cases that AI escalates) preserve most jobs. Operators that do not transition will lose market share to those that do. Jamaica's SAFE Task Force and Decent Work Recognition Programme are early frameworks for keeping the sector regulated through the transition.
Run a workforce exposure audit, stand up AI literacy programmes for highest-exposure staff, fund a regional skills transition mechanism through CARICOM and the Caribbean Development Bank, retrofit tertiary curricula for AI augmentation, and publish a national AI labour transition policy this year. Jamaica's 2025 unemployment insurance benefit for BPO workers is the closest regional template.
Build the skills that AI cannot replace: domain expertise, critical reading, judgement under uncertainty, and the ability to direct AI tools rather than execute the same work by hand. The worker who learns to instruct the model earns materially more than the worker who does the same task without it. The opportunity is real for any worker with three to six months of disciplined practice.
The ILO-NASK 2025 Global Index assesses occupational exposure at the 6-digit ISCO-08 level across roughly 30,000 tasks, combining task-level data with expert input and AI model predictions. Exposure does not mean automatic job loss; the framework distinguishes high-risk automation from task-level transformation. Most exposed jobs will be transformed rather than replaced.
Editor's note

The Caribbean has lived through automation waves before, and the region's BPO sector grew because of one of them. AI is the first wave where the same worker who would be replaced can also be the one directing the replacement, if the literacy investment lands first. The transition window is this year. After 2027, the workers who were highest-earning in BPO and clerical roles will be either operating at three times their current productivity or competing for the half of the work that AI cannot yet do.

Caribbean AI Newsletter  /  June 2026

About Caribbean AI

Caribbean AI is the official directory of artificial intelligence companies, labs, and innovators in the Caribbean. We connect startups, enterprises, and researchers driving the region's AI growth.

For applied AI advisory and workforce transition support across the region, visit starappleai.org.

JamaicaTrinidad & TobagoBarbados GuyanaBahamasSaint Lucia GrenadaAntiguaDominica BelizeSurinameCuraçao Dominican RepublicSt VincentHaitiAruba
Explore the Directory
Next
Next

3-month Caribbean Claude AI study [exclusive insights]