Sustainable artificial intelligence in global health: balancing promise, equity, and environmental realities in low- and middle-income countries.
TL;DR
Funders and implementers should prioritize durable infrastructure, strengthened health information systems, rigorous evidence generation, environmentally responsible deployment, and local stewardship to achieve scalable, equitable, and sustainable AI in global health.
Abstract
Objective
Artificial intelligence (AI) adoption in global health informatics is accelerating, yet scaling, sustainability, equity, and environmental challenges limit impact, particularly in Low and Middle Income Countries (LMICs).
Materials And Methods
Drawing on experience from members of the American Medical Informatics Association Global Health Informatics and Climate, Health and Informatics Working Groups, we synthesized implementation, evaluation, sustainability, and governance considerations for AI in resource constrained health systems.
Results
We propose a framework integrating four components: Green AI necessity assessments; a One Digital Health systems lens; pragmatic, workflow integrated evaluation; and federated governance supporting locally led stewardship and cross institutional learning.
Discussion
Sustainable AI requires moving beyond short term pilots to address infrastructure, environmental costs, workflow integration, equity, and locally relevant evidence.
Conclusion
Funders and implementers should prioritize durable infrastructure, strengthened health information systems, rigorous evidence generation, environmentally responsible deployment, and local stewardship to achieve scalable, equitable, and sustainable AI in global health.