Generative AI and Indigenous Knowledge Systems: Opportunities and Risks for Cultural Preservation

Abstract

Generative artificial intelligence is increasingly capable of producing, translating, reconstructing, and organizing culturally significant text, speech, imagery, and oral narratives. These capabilities create new possibilities for documenting Indigenous languages, transmitting community knowledge, restoring fragmented archives, and improving intergenerational access to cultural heritage. At the same time, generative models can reproduce cultural inaccuracies, fabricate traditions, erase contextual distinctions, and reuse Indigenous knowledge without appropriate consent or attribution. Existing research has largely treated cultural preservation as a problem of digitization, language documentation, or technical accessibility rather than one of community-controlled generative representation.
This study addresses that gap by proposing a culturally governed framework in which generative AI is evaluated not only for computational performance but also for provenance fidelity, cultural contextuality, community authority, and knowledge-use restrictions.
The research conceptualizes Indigenous knowledge as relational and governed information rather than an unrestricted corpus available for conventional machine-learning extraction.