Capacity building for ethical use of artificial intelligence in health: protocol for a scoping review of training initiatives and gaps in Africa

Abstract

Introduction: Artificial intelligence (AI) is increasingly embedded in health systems, with applications spanning diagnostic imaging, clinical decision support, disease surveillance and health system planning. While international guidance frameworks outline principles for safe and ethical AI deployment, effective governance depends on the capacity of regulators, research ethics committees, policymakers, health technology assessment bodies and frontline health professionals to evaluate, supervise and implement AI tools in practice. Across African Union (AU) member states, training and capacity-building initiatives related to ethical AI use in health remain fragmented and unevenly documented. This scoping review aims to systematically map, characterise and synthesise evidence on training initiatives that support the ethical use and governance of AI in health across Africa. Methods and analysis: This scoping review will follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines and applies the Joanna Briggs Institute population–concept–context framework. We will search PubMed/MEDLINE, Scopus, Web of Science Core Collection, Embase, IEEE Xplore, SSRN and African Journals Online, alongside structured grey literature searches of African regional bodies (eg, AU, Africa Centers for Disease Control and Prevention, AUDA-NEPAD, WHO Regional Office for Africa), national regulatory authorities, ministries of health, professional councils and donor programme portals. Searches will cover 1 January 2015 to 31 December 2025 and will be limited to English-language materials. Eligible sources must describe identifiable training, education or capacity-building initiatives focused on ethical use and/or governance of AI in health, including software as a medical device and AI-enabled medical devices, delivered in or explicitly targeting at least one AU member state. Purely technical AI training without governance or clinical integration components will be excluded. Two reviewers will independently screen records and extract data. Findings will be synthesised using descriptive statistics and reflexive thematic analysis. Where feasible, stakeholder consultation will be undertaken to contextualise the results and refine the draft competency framework. Ethics and dissemination: The review will analyse publicly available documents and does not involve human participants; formal ethical approval is not required. Findings will be disseminated through a peer-reviewed open-access publication, a policy brief targeted at regional and national health governance stakeholders, and a public webinar. Search strategies, screening decisions and extraction tools will be deposited on the Open Science Framework (OSF) to enhance transparency.