Abstract
Maternal mortality remains a critical global public health issue, particularly in low- and middle-income settings where failures in surveillance, early diagnosis, and clinical decision making compromise obstetric care. In this context, the present study aimed to critically review the scientific literature on the use of artificial intelligence (AI) in gynecologic and obstetric health, focusing on the prevention of avoidable deaths and severe maternal events. This is a narrative literature review with a qualitative and exploratory-interpretative approach, conducted between May and June 2025 in relevant electronic databases, following structured axes for narrative reviews and well-defined eligibility criteria. A total of 403 records were identified, of which 17 studies met the criteria and were included in the analysis, supported by the Interface de R pour les Analyses Multidimensionnelles de Textes et de Questionnaires software. The results showed that AI has been used to predict obstetric risks such as pre-eclampsia, postpartum hemorrhage, and preterm birth through machine learning algorithms, neural networks and predictive models based on electronic health records and laboratory tests. Tools such as clinical decision support systems, portable devices, and mobile applications have also optimized care, particularly in regions with limited infrastructure. However, challenges remain concerning the validation of algorithms across diverse populations, the inclusion of sociodemographic variables, and ethical considerations. In conclusion, AI is a promising technology in obstetric care, with the potential to reduce maternal morbidity and mortality. Nevertheless, its implementation requires ethical guidelines, adequate professional training, and inclusive public policies that promote digital equity in maternal and child healthcare.