Detalhes do Documento

The persistent high incidence of human immunodeficiency virus (HIV) and other sexually transmitted infections (STIs) constitutes a critical public health challenge in Latin America, a region characterized by structural inequalities and epidemiological heterogeneity. Although artificial intelligence (AI) has demonstrated transformative potential for disease surveillance and diagnosis globally, its application within the Latin American context remains underexplored. This study aims to provide a comprehensive assessment of AI applications for HIV and STI detection and prevention in Latin America, identifying principal technological advances, mapping the intellectual structure of this emerging field, and critically evaluating implementation advantages and limitations. Employing an exploratory qualitative design, a bibliometric review coupled with thematic synthesis was conducted. Five electronic databases were searched for studies published between January 2015 and December 2025. Following PRISMA guidelines, a multi-stage selection process involving two independent reviewers identified 28 empirical studies for final inclusion. Bibliometric mapping using VOSviewer examined keyword co-occurrence and collaborative networks, while thematic analysis using NVivo 14 integrated deductive and inductive coding. Bibliometric analysis revealed a rapidly expanding research landscape dominated by Brazil, Mexico, and Colombia. Thematic synthesis demonstrated that machine learning models, particularly ensemble methods such as XGBoost, consistently outperformed traditional approaches in predictive accuracy, achieving metrics exceeding 85% in regional validation studies. However, persistent challenges emerged: algorithmic bias from unrepresentative training data, fragmented health information systems, ethical concerns regarding data privacy, and a significant implementation gap between technical validation and public health impact. Artificial intelligence offers substantial promise for enhancing HIV and STI prevention in Latin America through improved diagnostic precision and targeted interventions. Realizing this potential requires sustained investment in digital infrastructure, interdisciplinary collaboration, and robust ethical governance frameworks to ensure equity and protect vulnerable populations.