Data and local initiatives in Africa

Several obstacles are slowing the rollout of locally developed AI systems powered by continent-specific data, and the integration of African languages spoken by rural populations: among these are limited access to high-calibre servers and a lack of digitised, large-scale datasets that can be used by AI methods.

Agricultural production at the heart of AI development

Among these new AI-driven development strategies, African agriculture — together with health and education — is a high-priority sector. 

Agriculture, health and education, three high-priority sectors for an IA serving Africa.

‘Digital technology, and more recently AI, have already transformed the African agricultural sector. New agri-tech start-ups  and research institutes are some of the stakeholders contributing to more effective uses of AI to improve agricultural production in Africa, with the aim of ensuring its sustainable development,’ explains Paulin Melatagia Yonta, a computer science researcher at the University of Yaoundé I. One of the projects he coordinates, supported by Cirad, is testing the ability of neural networks to simulate cocoa tree growth in Cameroon.

The Pixfruit project, developed by Julien Sarron, a Cirad researcher (HortSys Research Unit) at the Convergence Institute #DigitAg, also illustrates this trend. Based on drone imagery and AI (machine learning, neural networks), the tool had a dual objective: to map land use and estimate mango yields at the individual tree level in Senegal, within and across orchards, all the way to a smartphone application. ‘Today, we no longer have a partner start-up for Pixfruit, so we brought the tool in-house. Our research is now focused on further developing our expertise and applying it to other crops such as citrus, coffee and cocoa,’ explains Julien Sarron. Likewise, since late 2020, the team has been conducting specialised research on Corsican clementines with Laurent Julhia, an INRAE researcher at the AGAP Institute.

The rise of private companies 

Across the African continent, several private companies have already developed AI-based solutions to monitor plantations and optimise practices. The Morshida solution (from the Moroccan company DeppLeaf) detects plant diseases in real time; Agrix Tech, an application created in Cameroon, identifies plant diseases and proposes appropriate treatments; and Tolbi, a Senegalese agri-tech company, develops solutions to predict yields, calculate input requirements and anticipate risks linked to climate problems, diseases and pests.

The rise of academic research

To complete this overview, the DAAfrica Datascience for Agriculture in Africa workshop, organised by Paulin Melatagia Yonta from the University of Yaoundé I and Mathieu Roche from Cirad, recently focused on AI and data science approaches in agriculture. #DigitAg also supports work on predicting food security indicators in Africa from heterogeneous datasets using deep learning methods. Other research projects take the form of PhD theses in AI, with applications in epidemiological monitoring for animal and plant health, conducted through partnerships between France (Cirad), Kenya (Strathmore University) and Morocco (Sultan Moulay Slimane University).

In this dynamic context, it is important to highlight the Transforming Food Systems and Agriculture through Research in Partnership with Africa (TSARA) initiative launched in 2022 by 22 African and French member institutions, including Cirad and INRAE. One of the programme’s objectives is to explore how AI —particularly generative AI— can provide advice to farmers, advisers and policymakers with the goal of ensuring food security.