AI-Driven Mobile Framework for Preserving Tribal Knowledge Systems Integration of ASR, NLP, and GIS under Data Sovereignty Principles
- Posted
- Server
- Preprints.org
- DOI
- 10.20944/preprints202510.1116.v1
Oral traditions, ecological practices, and customary rules are all ingrained in the rich cultural history and knowledge systems of tribal societies. However, globalization, displacement, and a decline in intergenerational transmission are posing a growing danger to these knowledge systems. In order to overcome these obstacles, this study suggests a mobile framework powered by artificial intelligence (AI) that combines Geographic Information Systems (GIS), Natural Language Processing (NLP), Automatic Speech Recognition (ASR), and Machine Translation (MT) for documentation and preservation. The methodology uses a mixed-methods approach, integrating AI-based simulations with ethnographic research. Audio-visual recordings of ecological wisdom and oral histories are made using mobile devices and annotated using Praat and ELAN. While NLP and MT use Marian NMT and Hugging Face models for translation (assessed using BLEU and METEOR), the ASR pipeline uses Kaldi, ESPnet, and wav2vec 2.0 for transcription (measured with Word Error Rate). Cultural and ecological sites are documented by GIS mapping using QGIS and ArcGIS. Data is kept in community-controlled Mukurtu archives, and ethical considerations are incorporated through Indigenous Data Sovereignty (IDS) and the CARE Principles. Initial results indicate 90% accuracy in GIS validation, BLEU scores ranging from 24 to 31, and WER between 18 and 22%. Although privacy concerns are raised, community surveys show that mobile-AI tools are well accepted (easy of use = 4.3/5). In addition to providing scalable technological solutions with policy importance in education, e-governance, and cultural sustainability, this study offers a socio-technical and ethical framework for the preservation of tribal knowledge.