Artificial intelligence for sexual, reproductive and maternal health in Latin America and the Caribbean: a scoping review.

BACKGROUND

Artificial intelligence (AI) holds considerable promise for strengthening sexual, reproductive and maternal health (SRMH) by enhancing diagnosis, optimising service delivery and expanding access to information. In Latin America and the Caribbean (LAC), however, the scope, focus and maturity of AI applications in SRMH remain poorly described.

OBJECTIVE

To identify, map and analyse existing applications of AI in SRMH priority services in LAC, characterising thematic areas, target populations, and types of AI tools, whilst highlighting gaps and future research needs.

METHODS

We conducted a scoping review guided by the Arksey and O'Malley framework and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA-ScR). Searches were performed in PubMed, SciELO, Cochrane and LILACS up to August 2023, complemented by targeted Google searches and snowballing. We included records reporting AI applications in SRMH services in LAC and extracted data on setting, population, SRMH domain, AI techniques and implementation stage.

RESULTS

A total of 1,518 records were identified, of which 143 met the inclusion criteria. Most were published between 2020 and 2023 and originated from Mexico, Colombia, Peru, Brazil and Argentina. Over half were peer-reviewed articles, with additional theses and web-based reports. Applications concentrated on prenatal, childbirth and postnatal care (36%) and reproductive organ cancers (31%), with far fewer initiatives addressing sexual health, contraception, gender-based violence, sexual satisfaction or counselling. Machine learning methods predominated (52%), followed by deep learning (41%). Almost half of initiatives (48%) were exploratory projects, 17% implemented tools without outcome data and 35% reported performance in real-world contexts.

CONCLUSIONS

AI applications in SRMH in LAC are expanding but remain thematically narrow, population-selective and predominantly exploratory, highlighting the need for more diverse, rigorously evaluated and equity-oriented tools.