Global RETFound at the African Ophthalmology Council Congress
A year on from the launch of Global RETFound, an initiative to develop the first globally representative artificial intelligence (AI) foundation model in medicine, the number of collaborating study groups continues to grow across the six populated continents. Our colleague Fiston Kitema recently presented the project at the African Ophthalmology Council (AOC) Congress in Cape Town, and reports on the learnings from regional health leaders on their priorities and the challenges they face in harnessing data for medical AI.
I am grateful for the support that enabled me to attend and present at the African Ophthalmology Council (AOC) Congress in Cape Town in August. The congress brought together ophthalmologists, researchers, eye care managers and policy leaders from across the continent under the theme “Ophthalmology from Cape to Cairo – Unifying, Innovating, Defining Standards.” For me, it was more than a scientific meeting: it was an opportunity to move the conversation about fair medical artificial intelligence from principle to practical African participation.
The purpose was to present the mission of the Global RETFound Consortium, listen to the priorities of African eye care leaders, and identify institutions that could contribute to, and benefit from, a more representative medical foundation model. The model is being developed by the Pearse Keane medical AI lab at UCL Institute of Ophthalmology and Moorfields Eye Hospital, the National University of Singapore, and the Chinese University of Hong Kong, in collaboration with global research partners.
Africa remains under-represented in many medical AI datasets, despite carrying a substantial burden of avoidable vision loss and encompassing enormous genetic, demographic, clinical and health system diversity. If African population data is absent during model development, tools may perform less reliably and may not benefit the continent that needs this technology the most.
The AOC 2026 was therefore an important convening point. Its pan-African audience made it possible to speak directly with people who understand their national datasets, clinical workflows, compute infrastructure, and ethical frameworks. The aim was not simply to recruit data contributors, it was to build relationships, surface barriers early and encourage African partners to contribute to the development of AI, shape governance, and inform its future implementation.
Calling for African participation
I presented at the Technology for Eye Care session on the theme “ Developing medical AI in Africa through the Global RETFound Consortium”. It drew on our recent paper, “Addressing the African ophthalmic AI gap”, which argues that contextualised AI (trained or adapted for local conditions), combined with accessible imaging technologies, could extend eye screening and care to underserved communities, but only if limitations in data, infrastructure, validation and local capacity are addressed.
The original RETFound model, developed by researchers at UCL Institute of Ophthalmology, Moorfields Eye Hospital and the INSIGHT Eye and Oculomics Health Data Research Hub at Moorfields, demonstrated that a foundation model could be adapted to multiple ocular and systemic disease tasks using little labelled data. Although most of the training data reflected London's diverse population, it still doesn't capture the range of genetic and demographic variation found worldwide.
Global RETFound responds to this limitation by bringing together international partners to build a more geographically and demographically representative foundation model from retinal images using data from across the world. The consortium’s ambition is important for Africa for two reasons. First, broader representation can support more robust and equitable AI model development and evaluation across populations, devices and care settings. Second, the collaboration can create a platform for African led research provided participation extends beyond data sharing.
Immediately after my presentation, I had the privilege of inviting Professor Ogugua Okonkwo to the stage to highlight the need for African participation in the Global RETFound Consortium, and share firsthand the experience of being involved. Professor Okonkwo is one of Africa’s leading retinal surgeons and a Professor of Ophthalmology whose advocacy reinforced the importance of African institutions joining the consortium. That moment captured the value of trusted continental leadership in demonstrating that global collaboration becomes meaningful when the message is carried and shaped by leaders from the region.
Barriers to data management
The most valuable part of the congress came after the formal presentation. In individual discussions, and during focused meetings with researchers from across Africa, three recurring challenges emerged:
Data storage and management: Several colleagues were uncertain about the volume, location or structure of retinal image archives at their institutions. Limited server capacity, fragmented storage, inconsistent metadata, and a shortage of data management expertise can make participation difficult, even when imaging has been performed for years.
Limited access to ophthalmic imaging: Colleagues described countries and services with very few or no retinal imaging devices at all. Short-term outreach programmes may deliver valuable care, but when equipment, image archives and local maintenance capacity do not remain, they do not enable a durable data management infrastructure suitable for research, longitudinal care or AI development.
Limited protected time for governance: In many institutions, ophthalmologists carry clinical, administrative and research responsibilities simultaneously. Preparing ethics applications, data inventories, agreements and governance documentation can therefore stall, not because of a lack of interest, but because the process competes with urgent clinical work and may have little local administrative support.
Reflection and next steps
There is hope that strong African participation will result in African-led scientific outputs and strengthened local capacity. This could include mentorship in retinal data curation, secure data management, AI methods and model evaluation; opportunities for early-career researchers to lead analyses and publications; and support for institutions to develop sustainable imaging and research infrastructure. Partnerships with governments, universities, hospitals, funders and imaging device manufacturers may also help address settings where insufficient equipment currently prevents both routine retinal imaging and participation in AI research.
The consortium's goal is not only building a larger, inclusive medical foundation model, but also to create a reusable global research resource that supports locally relevant innovation, while reducing rather than reproducing existing inequalities in medical AI.
The conversations initiated at AOC 2026 provide a solid foundation for transformative work in Africa. The Global RETFound team stands ready to support and engage with any consortium member or any institution willing to join this impactful work.
For further information or to discuss joining the Global RETFound Consortium, please contact f.kitema@ucl.ac.uk
About Fiston Kitema
Dr Gatera Fiston Kitema is a vision science practitioner at Moorfields Eye Hospital and a visiting research fellow at the UCL Institute of Ophthalmology with a PhD in global eye health. His interests include AI applied to ophthalmic imaging, clinical research, and preventive ophthalmology. He contributes to the Global RETFound consortium, supporting international scientific engagement.
Acknowledgements
Special thanks to the Pearse Keane Medical AI Lab at the UCL Institute of Ophthalmology for funding the visit, and to the management of the Moorfields Medical Imaging and Ultrasound Department for their support.
Further reading
