<![CDATA[Indigenous medicinal plant knowledge constitutes a crucial component of ecolinguistic systems, as it is embedded in linguistic expressions that reflect ecological relationships, healing practices, and cultural values. However, this knowledge is increasingly threatened by language shift and insufficient documentation, particularly within low-resource indigenous communities. This study develops an AI-based ecolinguistic framework to systematically document and represent Nias ethnomedicinal knowledge by integrating ethnobotanical field data with culturally grounded artificial intelligence approaches. Qualitative data were obtained through semi-structured interviews with traditional Nias healers, resulting in the identification of fifteen commonly used medicinal plant species. To assess cultural salience and communal consensus, the study applied the Relative Frequency of Citation (RFC) index. The quantitative findings reveal an uneven distribution of cultural prominence among the documented species. Notably, Gundre and Mbulu Nazalöu emerged as the most frequently cited plants (FC = 14; RFC = 0.93 each), indicating their central role within the Nias ethnomedical knowledge system. The documented knowledge was subsequently structured using a Knowledge Graph model and enhanced through a Retrieval-Augmented Generation (RAG) architecture to enable contextualized, culturally sensitive knowledge representation. The proposed framework demonstrates how artificial intelligence can support the preservation, organization, and revitalization of endangered indigenous medicinal knowledge while maintaining its ecolinguistic integrity.]]>
Copyrights © 2026
