Large language models (LLMs) have been proposed to address global health inequity by providing accessible and high-quality health care, particularly in low- and middle-income countries (LMICs). However, despite the early enthusiasm following the release of GPT, development and deployment of LLMs have remained heavily concentrated in high-income countries (HIC), raising concerns that such technology may worsen existing health disparities instead of alleviating them. The most recent LLMs, which include features such as lower cost, and open-source framework, show promise in rebalancing LLMs’ benefits worldwide. In this viewpoint, we examine the current challenges and imbalance in LLM deployment across global regions, identify the key barriers to adoption in LMICs, assess current LLMs’ advances and the new opportunities they bring to global health equity. We also propose a five-dimensional roadmap—focusing on people, products, platforms, processes, and policies—to advance LLMs’ equitable adoption in LMIC and improve inclusive progress in global health. Funding: National Key R&D Program (Grant No: 2022YFC2502800); National Natural Science Fund of China (Grant No: 82388101); Beijing Natural Science Foundation (Grant No: IS23096).
