BEYOND PREDICTION: A RESEARCH AGENDA FOR ADVANCED AI IN AFRICAN SOIL INFORMATION SYSTEMS

Valantine Asong Asong Tellen , Chrow Khurshid, Luca Brocca, Therese Ncheuveu Nkwatoh, Kwame Twum Adjei, Godswill Azinwie Asongwe

Digital soil mapping has produced the first continent-wide baseline estimates of soil properties in Africa, yet existing soil information systems remain fragmented, poorly standardized, and largely confined to static offline spatial prediction. Their ability to capture near-real-time soil conditions or provide timely, field-level decision support is therefore limited. This paper argues that the next step is not simply applying artificial intelligence to improve mapping accuracy, but adopting Advanced AI as the foundation for adaptive Soil Digital Twins that continuously integrate new observations and support evidence-based management. It proposes a ten-pathway research agenda structured around four interconnected pillars: (1) Unlocking legacy archives using Natural Language Processing and Computer Vision to convert unstructured historical records into machine-actionable, AI-ready data; (2) Engineering dynamic systems through Online Learning, Generative AI, and Physics-Informed Neural Networks to continuously assimilate streaming telemetry; (3) Delivering prescriptive intelligence using multi-objective recommendation engines, human-centered Explainable AI (XAI), and Human-in-the-Loop expert validation; and (4) Enforcing sovereign governance via Federated Learning, FAIR and CARE-aligned autonomous stewardship. To operationalize this vision, it calls for a Pan-African AI for Soil Health Consortium to establish sovereign compute hubs, secure sustainable financing, and cultivate a home-grown talent pipeline. This agenda provides a strategic blueprint for transforming fragmented soil observations into a dynamic, trusted public infrastructure that addresses land degradation and food insecurity, while supporting climate resilience and sustainable agriculture across the continent.

Keywords: Artificial Intelligence (AI), Soil Digital Twins, soil information systems (SIS), data sovereignty, Federated Learning, Natural Language Processing (NLP), Explainable AI (XAI), Africa.