An AI-Driven Workflow for Preserving Traditional Malay Music Videos: A Case Study in Cultural Heritage Enhancement
Mohd Izani, M. Gabr, H. Harumaini, Khoi Trinh, Akhmed Kaleel, A. Assad · 2025
Preserving cultural heritage like traditional arts and traditional music becomes increasingly challenging during the digital age. We presented a workflow that uses generative AI to preserve traditional Malay music in video format with an emphasis on cultural integrity through photorealistic video generation. Our enhanced framework uses Low-Rank Adaptation (LoRA) combined with Splitter.ai-based vocal isolation and Kling.ai motion synchronization along with other artificial intelligence techniques to completely process and enhance audiovisual elements. We started our research design by deploying generative AI tools throughout a production process starting from concept development, facial synthesis, and training alongside motion-lip synchronization before post-production to ensure visual harmony without sacrificing storytelling elements and technical precision. The workflow consists of expert validation points, loop refinement linked to cultural accuracy loss, and computational performance. The end results show that generative AI techniques create successful links between traditional artistic forms with the present multimedia formats to preserve traditional Malay music. This research introduces a workflow that allows reproducible AI-assisted preservation methods which give opportunities to heritage experts and authorities helpful guidance to integrate technical innovations with cultural integrity.