About the Project
AFRICAI-RI – Bridging the AI divide through pan-African and European collaboration
AI holds great promise to improve access to healthcare and medical imaging throughout Africa, notably in Sub-Saharan Africa (SSA), a region highly afflicted by infectious diseases such as tuberculosis (TB) and pneumonia (PNA). These diseases disproportionately affect adults and children living in poverty, exacerbated by challenges such as undernutrition, overcrowding, and high co-infection rates with HIV. Moreover, under-diagnosis is a major healthcare problem across SSA, with studies indicating TB remains undetected in about half of the individuals using passive case finding methods and in over 90% under active screening approaches.
While chest X-rays (CXR) and chest ultrasounds (CUS) offer accessible imaging modalities in SSA, the region suffers from a severe shortage of clinicians skilled in interpreting such scans for respiratory disease detection. With some areas having only one radiologist per hundreds of thousands of residents, the need for innovative diagnostic solutions is urgent. AI offers significant promise in meeting this critical clinical need by enabling automated image interpretation in under-served SSA regions with high rates of TB, PNA and HIV. However, the lack of open African repositories providing secure access to large imaging collections for training the AI tools has significantly hindered the much-need AI development in SSA.
A Vision for Africa-Driven Innovation
AFRICAI-RI’s goal is to establish the very first federated, multi-institution imaging data infrastructure for Africa, addressing a critical gap in the continent’s biomedical resources. The consortium will leverage extensive international expertise to adapt successful imaging solutions to fit the unique requirements and diverse environments within SSA. To this end, we will employ a highly inclusive, co-creation approach, with the active involvement of 10 centres from SSA, representing a spectrum of geographical, cultural and healthcare contexts.
The Challenge
While AI is transforming healthcare worldwide, most imaging AI tools are developed in high-income countries and trained on data that does not reflect African populations. African hospitals frequently conduct X-ray and/or ultrasound scans, yet lack secure infrastructures to store and analyse imaging data or collaborate across borders.
This limits:
- Access to high-quality African datasets
- Development of African-adapted AI tools
- Opportunities for young African scientists
- Research on diseases that disproportionately affect the region
Our Solution
The First Pan-African Imaging Research Infrastructure
AFRICAI-RI is creating:
- A network of secure, interoperable data nodes across ten African institutions
- A central support hub for onboarding new sites
- Advanced tools for data curation, quality checks and metadata management
- A federated learning platform enabling AI development without transferring sensitive data
This approach ensures privacy, trust, and local ownership.
Expected Results
By the end of the project, AFRICAI-RI will have:
Established a sustainable imaging infrastructure reaching at least 13 data nodes
More than 76,000 X-ray and ultrasound images curated
Developed AI tools tailored to African clinical needs
Enabled new clinical research across the continent
Delivered a large training programme for SSA partner researchers, clinicians and students
Supported real-world AI testing in multiple hospitals
Produced long-term sustainability and governance plans