Generative AI Technology for Safeguarding Intangible Cultural Heritage: A Systematic Review

Abstract

Generative AI, as an advanced technology, has demonstrated significant application potential in the field of the Safeguarding of Intangible Cultural Heritage (ICH) in recent years. This study employs a systematic review method to comprehensively review and analyze the current status of generative AI technologies in the Safeguarding of ICH. The study focuses on core generative AI technologies such as Generative Adversarial Networks, diffusion models, neural style transfer, autoregressive models, and variational autoencoders, as well as their practical applications in the field of ICH safeguarding. The results show that GANs and diffusion models dominate among all the technological types. Overall, generative AI technologies show significant potential in promoting the digital innovation and generation of ICH, fostering cross-cultural integration, and modernizing traditional arts. However, challenges remain, such as insufficient data samples, hardware limitations, and the inadequate capture of cultural connotations. Future research needs to further optimize algorithms, enhance cultural semantic understanding, and involve ICH inheritors to ensure that technological development is grounded in cultural contexts and aligns with cultural authenticity, thereby better promoting the sustainable safeguarding and transmission of ICH.