From the Editorial Office
The launch issue of JCAIS places the Caribbean at the centre of some of the most consequential conversations in contemporary AI: how agentic systems deliberate and decide; how world-scale foundation models can be appropriated and made local; how generative AI intersects with culture, biodiversity, and economic survival. These four papers are not peripheral responses to a field developed elsewhere — they are original contributions that expand the frontier. An experiment with 100 AI agents debating Caribbean national identity. A protein-structure analysis of Jamaican ethnomedicine. A hard economic reckoning with what AI music means for Reggae. A framework for nations to build their own language models. Each paper, in its own way, is a report on the future.
Table of Contents
Caribbean AI Research Initiative
Abstract
This paper presents a novel experimental framework for deploying large-scale agentic AI communities to deliberate on contested cultural and political questions. We report on a controlled experiment in which 100 autonomous AI agents — each configured with distinct demographic, political, and cultural personas drawn from Jamaican civil society — deliberated over multiple structured rounds on the question of whether reggae icon Bob Marley should be formally designated a National Hero of Jamaica. Using a multi-round argumentation and coalition-formation protocol, agents updated positions, formed alliances, and produced emergent consensus dynamics that we analyse against established models of political deliberation theory. We find that agentic communities can generate contextually rich, culturally grounded deliberative outputs and demonstrate measurable sensitivity to persona-level political context — while also exhibiting systematic biases requiring mitigation. The study offers a methodological foundation for AI-assisted civic consultation, policy simulation, and the computational study of political discourse in Caribbean societies.
Faculty of Medical Sciences, University of the West Indies, Mona
Abstract
This study applies DeepMind’s AlphaFold protein structure prediction system — one of the landmark world-model achievements in modern AI — to investigate the pharmacological basis of Jamaican traditional herbal remedies. We selected fifteen widely-used Jamaican ethnomedicinal plants, including soursop (Annona muricata), fever grass (Cymbopogon citratus), and cerasee (Momordica charantia), and used AlphaFold to generate three-dimensional protein structure predictions enabling in-silico molecular docking analysis against known disease-relevant receptors. Our analysis identifies plausible structural mechanisms for several remedies whose efficacy has been empirically observed but not previously explained at the molecular level, including statistically significant anti-inflammatory and anti-diabetic binding affinities. We discuss the implications of AI-driven structural biology for validating traditional pharmacopeia in resource-limited settings and argue that AlphaFold represents a transformative equaliser for drug discovery anchored in Caribbean biodiversity and indigenous knowledge systems.
Department of Economics, University of Technology Jamaica
Abstract
This paper develops a quantitative framework for estimating the economic displacement risk that generative AI music systems pose to Jamaica’s reggae industry — one of the country’s most significant cultural exports and a recognised UNESCO Intangible Cultural Heritage. Drawing on industry revenue data (2015–2025), streaming platform market share analysis, and scenario modelling calibrated against adoption curves observed in analogous creative-sector disruptions, we model potential revenue losses across three AI adoption trajectories over a five-year horizon (2026–2031). Our central estimate projects cumulative revenue displacement of USD 38–115 million, concentrated in session musicianship, music publishing royalties, and mid-tier artist income. We further analyse the cultural heritage dimension of reggae’s economic value as both a vulnerability and a potential legal protection mechanism, and propose a policy response package — including mandatory AI-music labelling, a proposed Caribbean Creative Industries AI Resilience Fund, and preferential streaming royalty structures for verified human artists — to protect Jamaica’s most significant cultural export in the generative AI era.
Caribbean AI Research Initiative
Abstract
This paper proposes a replicable framework for small and medium-sized nations to develop sovereign national-level large language models (NatLLMs) without the computational infrastructure and proprietary data access typically assumed by frontier AI development. The framework has two core components: first, a structured fine-tuning process for open-source foundation models — including LLaMA, Mistral, and Falcon architectures — applied to nationally curated text corpora; and second, a methodology for constructing digital twin training datasets comprising structured synthetic representations of national systems including legal and regulatory frameworks, public service processes, geographic and demographic data, and cultural artefacts. We operationalise the framework with a reference design for a Caribbean-state NatLLM, specifying data curation protocols, compute strategies, governance structures, and evaluation benchmarks appropriate for small-state deployment. We argue that the combination of open-source foundation models and digital twin data represents the most viable near-term pathway to AI sovereignty for nations that cannot compete for frontier compute, and present the framework as a public-domain contribution deployable by a small technical team with regional cloud infrastructure access.