AI without borders: reflexive thematic analysis of artificial intelligence justice in global health

Artificial intelligence (AI) is widely expected to mitigate global health inequities. However, current research mainly focuses on downstream issues such as algorithmic bias and data limitations and neglects upstream power structures and historical structural root causes. This study examines how AI deployment in global health may reproduce and reinforce structural inequities. A reflexive thematic analysis was conducted base on systematic searches of PubMed, Scopus, Embase, and Web of Science from database inception to April 2026. The search strategies combined three conceptual domains: “AI”, “global health” and “justice”. Supplementary searches were performed through reference list screening, with final inclusion determined by relevance to the research theme. After screening a total of 554 records, 37 studies were ultimately included in the qualitative analysis. A comprehensive justice framework was constructed to identify how AI-related justice dilemmas emerged across global health systems, with the Global South and resource-limited contexts serving as critical sites where global power asymmetries became especially visible. The findings reveal that justice is not limited to the distribution of resources but also encompasses multiple dimensions, including procedural, historical and structural, epistemic, and reparative justice. AI justice reflects global power relations; effective equity transformation requires examining and reconstructing power structures in which AI is embedded. Whether AI ultimately bridges or widens global health inequities depends not on algorithmic sophistication but on the willingness to redistribute the power that shapes its design, deployment, and governance.