INTELIGÊNCIA ARTIFICIAL NA ATENÇÃO PRIMÁRIA À SAÚDE BRASILEIRA: POTENCIALIDADES, RISCOS ÉTICOS E IMPLICAÇÕES PARA A EQUIDADE NO SUS

The incorporation of artificial intelligence (AI) into health systems has expanded opportunities for clinical decision support, risk stratification, task automation, epidemiological surveillance and care organization. However, in Primary Health Care (PHC), these applications operate in settings marked by territorial, racial, socioeconomic and digital inequalities, which may transform efficiency gains into new forms of exclusion. This study critically analyzed the potential benefits, ethical risks and equity implications arising from the use of AI in Brazilian PHC and the Unified Health System (SUS). An integrative review covering 2016–2026 was conducted using national and international scientific literature and regulatory documents on digital health, data protection and algorithmic governance. The synthesis addressed potential applications and benefits; sources of bias and inequality; data protection, transparency and accountability; and governance requirements for adoption in the SUS. AI may support early risk identification, reduce administrative workload, enhance teams’ analytical capacity and improve care coordination. Nevertheless, incomplete data, underrepresentation of population groups, inappropriate target variables, limited explainability, technological dependence and lack of subgroup monitoring may reproduce or intensify inequities. Ethical adoption in PHC therefore requires local validation, human oversight, continuous assessment of performance and equity, social participation, data protection and institutional accountability. In the SUS, technological innovation must remain subordinate to universality, comprehensiveness and equity and cannot replace structural investment, care relationships or public governance capacity.