IRIS Africa – Intelligent and Responsible Innovation with GenAI for Societal impact

Sub-Saharan Africa is confronted with interconnected challenges in agriculture, healthcare, and digitalization that affect sustainable development, resilience, and inclusion. These structural weaknesses require transformative approaches able to respond to urgent needs while unlocking long-term opportunities for growth and innovation. Generative AI represents a highly promising market with strong.

capacity for growth, but existing tools are often developed and deployed by Western actors, embedding biases that overlook African realities and reinforcing euro-centric perspectives. The IRIS project seeks to unlock the transformative potential of Generative AI (GenAI) for African markets, empowering communities through advanced computing solutions in agriculture and health. IRIS will employ Authentic Intelligence to customize and adapt two AI tools: Large Language Models (LLMs) and deep learning-based risk-classification algorithms. ZEF Health and its partner KNUST in Kumasi, Ghana are responsible for the co-design, implementation, evaluation, and deployment of a use case in health: continuous glucose monitoring (CGM) for improved management of diabetes mellitus in Ghana.

Methodology

This implementation science project takes a mixed-methods approach. In the test phase, 20 participants with diabetes mellitus and their physicians will gain 1-month experience in using CGM for diabetes management. In the co-design phase, the test data of glucose control and the participants’ lived experiences will inform the design of AI-supported communication of the CGM outputs to the patients and clinicians – either in local language text, or text-to-speech voice messages, or iconographic illustrations. In the trial phase, 150 participants with diabetes mellitus will be randomized to the AI-supported CGM versus conventional CGM to evaluate the effects on time in range, long-term glucose control, and other relevant parameters over 3 months. In the deployment phase, after comprehensive mapping, we will involve interest groups (MedTech, healthcare, health insurances, legislation, etc.) to facilitate the uptake of the AI-supported CGM to improve diabetes management in Ghana.