Inteligência artificial aplicada à prevenção, ao diagnóstico e ao cuidado de pessoas vivendo com HIV: desafios éticos, evidências científicas e perspectivas para a enfermagem brasileira

This article analyzed scientific evidence on the application of Artificial Intelligence (AI) to HIV prevention, diagnosis, and care, emphasizing ethical challenges and perspectives for Brazilian nursing. An integrative literature review was conducted using PubMed/MEDLINE, SciELO, the Virtual Health Library/LILACS, and DOI tracking mechanisms, including publications from 2015 to August 2026 in Portuguese, English, or Spanish. Twenty-five studies with verifiable DOI were selected, covering machine-learning applications for risk stratification and pre-exposure prophylaxis indication, automated interpretation of rapid tests, prediction of adherence, disengagement and viral suppression, counseling support, and algorithmic equity. Findings showed that AI can broaden the identification of priority populations, support clinical decisions, improve diagnostic interpretation, and enable personalized interventions during therapeutic follow-up. However, data bias, model opacity, privacy, stigma, information security, and technological dependence require governance, local validation, and professional oversight. It was concluded that Brazilian nursing can play a strategic role in the responsible implementation of these technologies by combining clinical judgment, health education, data protection, and person-centered care.