Culturally-Responsive AI Assessment Systems for Ohangla Music: A Literature Report

Abstract

The purpose of this article is to address the systematic bias in artificial intelligence (AI) music assessment systems that marginalise Indigenous African musical forms, specifically Ohangla music from Kenya’s Luo community. Current AI systems, designed around Western-centric musical paradigms, fail to recognise Ohangla’s complex polyrhythms and multifaceted cultural functions, resulting in misclassification and epistemological erasure of Indigenous musical knowledge. This study employs a systematic literature review synthesising thirty sources across ethnomusicology, AI bias research, technical music AI development, and Indigenous data sovereignty frameworks. It analyses Ohangla’s cultural significance, documents bias mechanisms in existing AI systems, evaluates technical approaches to music assessment, and examines ethical protocols for Indigenous music AI development. Findings reveal three primary mechanisms of exclusion: (1) training data bias that omits Indigenous music, (2) algorithmic assumptions privileging Western musical structures, and (3) evaluation criteria incompatible with community-centred assessment. Technical approaches succeed only in culturally homogeneous contexts, with African polyrhythms misclassified as computational “outliers” despite their mathematical sophistication. Ohangla’s role as a social technology promoting community bonding and cultural transmission challenges AI paradigms that prioritise technical optimisation over cultural preservation. The study concludes by proposing a culturally-responsive AI framework for Ohangla assessment that integrates Indigenous data sovereignty principles (OCAP®) with technical innovation. This framework provides methodological guidelines for community-centred AI development, advancing decolonising research methodologies and practical cultural preservation technology. It offers actionable recommendations for developers and policymakers to design AI systems that preserve rather than erase Indigenous cultural heritage, establishing a model for culturally-responsive AI in other Indigenous music contexts globally.