JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) has the potential to strengthen health systems in low- and middle-income countries (LMICs), but its development and deployment raise significant governance challenges. Most AI governance research remains focused on high-income settings, with limited empirical insights from LMIC health innovators.

Objective:

This study explores how researchers and innovators working on health AI in LMICs perceive and navigate governance challenges, and identifies opportunities to support more inclusive, effective, and context-sensitive AI governance systems.

Methods:

We conducted 30 semi-structured interviews with health AI developers, researchers, and implementation leads based in LMICs, using purposive sampling across academia, private sector, and multilateral organizations. Data were analyzed thematically using a constructivist grounded theory approach.

Results:

Participants highlighted a lack of clear and enforceable AI-specific regulation in LMICs, limited coordination between ministries and sectors, and weak integration of ethics into research and innovation processes. Infrastructure and capacity challenges were compounded by poor alignment between funding incentives and ethical standards. Participants emphasized the need for participatory, locally driven governance approaches that go beyond importing external standards.

Conclusions:

Strengthening AI governance in LMICs requires context-sensitive regulatory reform, improved public-private coordination, and support for locally led innovation. Policy frameworks must be grounded in local realities and supported by ethics-linked funding, capacity-building, and inclusive governance models.