AI for the People: How Haiti Can Use Technology to Build Resilience

Any honest conversation about AI in Haiti must begin with an honest conversation about Haiti. Not the Haiti of foreign crisis narratives, which flatten a complex and resilient nation into a single register of disaster and dysfunction, but the Haiti of the world's first successful slave rebellion, the country that inspired independence movements across the Americas, the place that produced Edwidge Danticat, Franketienne, and one of the richest oral literary traditions in the hemisphere. Understanding what AI can do for Haiti requires understanding what Haiti already is.

It is a country of 12 million people, the great majority of whom speak Haitian Creole as their primary language. It is a country with a history of extraordinary creativity and intellectual production operating under extreme resource constraints. It is a country whose diaspora, spread across Miami, Boston, Montreal, Paris, and beyond, has achieved at the highest levels in medicine, technology, law, and the arts. And it is a country that, for reasons rooted in history, geography, and international economic architecture, has been systematically underinvested in for generations.

AI does not fix any of that. But used well, by Haitian institutions and Haitian leaders, AI can make some of the most acute material problems that Haitians face more tractable. That is worth examining carefully and honestly.

The Creole Language Gap: AI's Most Important Failure in Haiti

Haitian Creole is not a dialect. It is a fully formed language with its own grammar, its own literature, its own orthography, and approximately 12 million speakers. It is the first language of virtually the entire Haitian population and the language in which Haitians conduct their daily lives, raise their children, worship, and tell their stories.

It is almost entirely absent from the training data of the world's major AI language models. This is not a minor gap. It means that AI-powered translation tools produce unreliable Creole output. It means that voice-to-text systems do not work for Creole speakers. It means that AI literacy tools, medical communication systems, agricultural advisory services, and public health information platforms are, for practical purposes, unavailable to Haitian Creole speakers. In a country where healthcare providers, disaster response agencies, and government services all struggle with communication barriers, this gap has direct human consequences.

The work of closing it is underway, but underfunded. The MIT-Haiti Initiative has been developing Creole educational technology and computing resources for over a decade. The Haitian Creole Wikipedia community has built a corpus of Creole text that, while still limited, provides a foundation for language model training. Researchers at the University of Florida, Carnegie Mellon, and several Haitian universities have published Creole NLP datasets and models. What is needed is the coordinated, sustained investment to bring these scattered efforts to the scale that makes Creole-capable AI genuinely useful.

CAIA has made Creole language AI one of its flagship research initiatives, working with diaspora academics, Haitian universities, and international research partners to build the dataset, the models, and the deployment infrastructure that can make AI work in Haiti's primary language. This is one of the clearest cases where the Caribbean AI community has both the capability and the obligation to do something the global AI industry will not prioritise on its own.

Earthquake Intelligence: Learning from 2010 and 2021

The 2010 earthquake that struck near Leogane, 25 kilometres west of Port-au-Prince, killed more than 200,000 people and left 1.5 million homeless. The 2021 earthquake in the southern peninsula, near Les Cayes, killed more than 2,200 people and destroyed tens of thousands of homes. Both events were on the Enriquillo-Plantain Garden fault system that runs through southern Haiti, a fault that seismologists have identified as carrying significant accumulated stress for potential future events.

Earthquake early warning systems cannot prevent earthquakes. But they can provide seconds to minutes of warning before shaking arrives at population centres, enough time for people to take protective positions, for trains and vehicles to slow, for industrial processes to shut down safely, and for emergency response systems to begin mobilising. Mexico City's Seismic Alert System, which uses AI pattern recognition across a network of sensors on the Pacific Coast to provide 60 to 90 seconds of warning before shaking reaches the city, has saved lives in multiple significant events. A comparable system for Haiti's fault systems is technically feasible and has been studied by regional seismology programmes.

Post-disaster, AI-powered satellite imagery analysis has transformed damage assessment. The Copernicus Emergency Management Service, which provides AI-assisted damage mapping for disasters globally, was deployed after both the 2010 and 2021 Haiti earthquakes. Systems that can classify building damage across thousands of structures in hours, rather than the days that ground surveys require, allow search and rescue resources to be directed where the probability of finding survivors is highest. Haiti needs permanent access to these tools and the training to operate them, not just emergency deployment when a disaster has already occurred.

Coffee, Cacao, and the Agricultural Comeback

Haiti was once the world's largest exporter of coffee. In the eighteenth century, Saint-Domingue, as Haiti was known under French colonial rule, produced more than half of the world's coffee supply and more than a third of its sugar. That agricultural heritage, buried under two centuries of deforestation, land degradation, and structural underinvestment, still has genuine residual potential.

The specialty coffee market has been rediscovering Haitian coffee. Varieties grown in the Massif du Nord, particularly the Haitian Blue Mountain variety grown at altitude in Laboule and Kenscoff, have attracted serious attention from specialty roasters in the United States and Europe. Cacao cultivation in the Grand'Anse and Nippes departments has been supported by organisations including the Cocoa Research Programme and several international development agencies, with Haitian cacao appearing in fine chocolate products from brands including Dandelion Chocolate. These are not marginal niche markets. They are premiums-priced, growing categories where Haitian origin can command significant price advantages.

Precision agriculture tools have demonstrated results in small-scale tropical agriculture that are directly applicable to Haitian conditions. Mobile soil analysis tools that can guide fertilisation decisions based on actual field conditions, rather than generalised recommendations, improve yields while reducing input costs. Pest identification AI that lets farmers photograph a leaf or fruit and receive a diagnosis with treatment recommendations reduces crop loss from identifiable diseases. Harvest timing models that integrate weather forecasting, market price data, and crop maturity indicators help farmers optimise the timing of their sales. These tools work on smartphones, require no permanent connectivity, and have been successfully deployed by agricultural NGOs in comparable smallholder environments in East Africa and Southeast Asia.

Partners in Health and the Healthcare AI Opportunity

Partners in Health has operated in Haiti since 1987. Through its Zanmi Lasante programme, it manages a network of hospitals and community health centres that provides care to millions of Haitians in the Central Plateau, the Artibonite, and the southern peninsula. It is, by a significant margin, the most capable healthcare institution operating in rural Haiti, and it holds one of the most extensive health datasets of any developing-country healthcare provider in the world.

PIH has been exploring AI-assisted clinical decision support with genuine seriousness. The organisation has the physician leadership, the informatics capacity, and the community health worker network to deploy AI diagnostic tools in ways that smaller organisations cannot. Chest X-ray AI that can screen for tuberculosis and pneumonia with accuracy comparable to a radiologist is already being used by PIH affiliates in Rwanda. Rapid diagnostic AI for malaria and other parasitic diseases has been deployed in West Africa with PIH support. Haiti's disease burden, which includes high rates of tuberculosis, malaria in rural areas, and the chronic disease burden that accompanies poverty, creates clear deployment targets for each of these tools.

The remittance economy is another AI-adjacent opportunity. Haiti receives more than $4 billion annually in diaspora remittances, representing more than a quarter of GDP. These transfers currently lose 5 to 10 percent to fees at every transaction. AI-powered mobile money platforms that compete on cost and convenience with Western Union and MoneyGram have dramatically reduced remittance costs in sub-Saharan Africa. Haiti's unbanked population, estimated at more than 80 percent of adults, could access formal financial services through AI-powered mobile platforms built on the remittance relationship that already exists between the diaspora and the homeland.

The Obligation of the Caribbean AI Community

Haiti is a Caribbean nation. Its people, its history, and its future are inseparable from the region's. The Caribbean AI community, including CAIA, has a specific obligation to Haiti that goes beyond the generic commitment to regional development. That obligation is to ensure that Haiti's AI future is not built by outsiders for Haitians, but by Haitians with the support of a Caribbean community that recognises its own stake in Haiti's resilience and prosperity.

That means investing in Haitian institutions, not routing resources through international intermediaries. It means building Creole language AI with Haitian linguists and communities, not importing English-language tools that do not work. It means listening to what Haitian technology leaders and entrepreneurs say they need, rather than deciding for them what AI should do in Haiti.

The Caribbean AI Association was founded on the conviction that the Caribbean must own its AI future. For that conviction to mean anything, it must include Haiti.