In this paper, the authors present a descriptive portfolio analysis examining the landscape of US National Institutes of Health (NIH)-funded artificial intelligence research in health science, with a particular focus on studies conducted in low- and middle-income countries (LMICs). The authors analysed 1,850 active NIH grants using the NIH RePORTER database, finding that only 97 (5.2%) were focused on LMICs, representing just 2.4% of the total $1.66 billion portfolio — a striking disparity that highlights the uneven global distribution of AI health research investment. Compared to high-income country studies, LMIC-focused grants placed greater emphasis on diagnostics and treatment, disease surveillance, health system optimisation, and telemedicine, reflecting the more pressing and immediate health challenges faced in these settings. The most commonly addressed health issues in LMIC-based research were non-communicable diseases, communicable diseases, and health system strengthening.

From a global health equity perspective, the implications are significant. While AI offers substantial potential to deliver cost-effective, scalable solutions in resource-constrained settings, the authors identify notable gaps in ethics, data governance, and public health communication within LMIC-focused research — areas that are critical for ensuring responsible and trustworthy AI deployment. The findings also underscore the need for greater local ownership, with only 31 awards made directly to LMIC-based investigators. The authors conclude that although NIH investment in AI-enabled global health research is growing, sustained attention to digital infrastructure, capacity strengthening, ethical frameworks, and meaningful international collaboration will be essential to ensure that the benefits of AI in health science are equitably shared across all populations.