Artificial intelligence (AI) is increasingly being applied in health research in Uganda, offering the potential to analyze data instantly and enhance diagnostic accuracy. However, the country currently lacks formal ethical guidelines for conducting AI-driven research. Evidence suggests that ethics committees are often under-resourced, undertrained, and lack the necessary expertise to adequately assess the unique risks posed by AI in health studies. In response, this study aimed to develop a structured framework to guide the ethical evaluation of AI system development in Uganda’s health research landscape. Between March and October 2024, this study utilized an exploratory qualitative design, engaging 35 stakeholders from two public universities in Uganda. Data were collected through in-depth interviews with twelve ethics committee members experienced in reviewing AI research protocols, six bioethicists, eight health researchers, and nine participants from AI development teams. The findings were analyzed using a thematic approach to identify key patterns and insights. Analysis of the data revealed six key themes: enhancing social value and equity; safeguarding the autonomy and safety of participants and end-users; addressing challenges related to data collection, access, and sharing; ensuring responsible data usage and minimizing unnecessary data retention; promoting ethical AI practices; and encouraging collaborative partnerships. Participants highlighted the potential of AI to advance health research but emphasized that its effective and safe application requires careful attention to ethical principles to protect both participants and end-users. Overall, participants believed that creating a structured guide for the ethical review of AI research could help mitigate potential risks associated with the use of AI tools in health studies. The study further recommends providing targeted training for ethics committees to enhance their understanding of the critical ethical considerations involved in developing responsible AI applications.
