Development of AI-Based Ethnobotanical Plant Recognition and Information System: The case of Ju/’hoansi San communities of Namibia

San community elders have been passing indigenous medicinal knowledge orally to the youth, and there is a risk of losing this knowledge. Researchers have been preserving this knowledge in digital formats, for example, the development of AI plant recognition and
Retrieval-augmented generation (RAG) systems have not clearly integrated Namibian endemic plants, especially those used for medicinal purposes by the San communities. This study aims to develop an AI system that will allow the San community to record, preserve, and share their medicinal practices. The objective is to collect ethnobotanical information and video materials with San communities in Namibia; to fine-tune an AI plant identification system integrated with a RAG model to identify endemic plant species and retrieve relevant medicinal videos accurately; to develop an ethnobotanical digital platform integrated with AI and RAG technologies; to test and validate the developed system in the field; and to implement a fully functional ethnobotanical digital system. The research approach will follow a pragmatic research paradigm that states that the research finds practical solutions to real-world problems. The study will be conducted with Ju/’hoansi San communities of Namibia. A convenient sampling approach will be used to select community members to collaborate in the study. The research will be implemented in the following five phases: collect ethnobotanical information and video materials, fine-tune AI and RAG systems, develop an AI ethnobotanical digital platform, Test and validate the developed system in the field, and implement and deploy a fully functional ethnobotanical digital system. The Expected outcome is the implementation of the system whereby a user will take or upload a picture of a plant leaf on the App, the AI plant recognition API will recognise the plant, and it will send the information to the retrieval Augmented Generation (RAG). It will retrieve relevant videos of a San knowledge holder explaining the medicinal use of the plant from the indigenous knowledge base.