Abstract Introduction Artificial intelligence (AI) has been proposed as a tool to expand access to medical care. The extent to which surgical AI research addresses the needs of low- and middle-income countries (LMICs) remains unclear. This scoping review synthesises primary research to assess where studies are being conducted and how far they progress towards clinical implementation. Methods A registered scoping review was performed in accordance with PRISMA-ScR guidelines (OSF.IO/9PV6A). PubMed, Scopus, and Web of Science were searched to 14/07/2025. Primary studies evaluating AI in LMIC surgical care were included. Extracted data were analysed to describe geographical distribution and the maturity of AI development, mapped to the OECD framework. Results From 2602 records, 474 studies met inclusion. Research was highly uneven with 64.6% of studies originating from China (n = 306). Only 9.5% (n = 45) originated from lower-middle income countries and 0.8% (n = 4) from low-income countries. Minimal evidence originated from Sub-Saharan Africa and Latin America. The evidence base was dominated by algorithm development and internal validation (n = 317, 66.9%). External validation was reported in 111 studies (23.4%), while clinical deployment was reported in 12 (2.5%). Translation beyond the development stage was particularly scarce in lower-income settings. Conclusions AI research in LMIC surgical care is expanding but unevenly distributed, with activity concentrated in higher-middle income Asian countries and striking gaps in low-income regions. Most work remains early stage, with little evidence of clinical use. Without investment in diverse data, infrastructure, and health system capacity, AI risks reinforcing global inequities rather than reducing them.
Valli et al. (2026) studied this question.