From “Classics” to “Radical Remix”: Aesthetic Reconstruction in AI-Generated Remix Videos

In recent years, the rapid growth of AIGC has brought a wave of AI-assisted remix videos to major online platforms. What began with relatively simple face-swapping and voice synthesis has gradually developed into dramatic rewritings of well-known film, television, and literary IP through generative AI. Once a marginal form of online parody, this type of content has become a visible part of contemporary digital culture. Its development can be understood as a shift from technological novelty to cultural deconstruction and, increasingly, to aesthetic reconstruction. In this article, aesthetic reconstruction is described as the method by which users use AI to depart from the original work's fixed narrative structure, character backgrounds and values, and reconstruct it into a novel form. Previous research has indicated that AI remixing modifies canonical texts by means of visual combination, change of narrative and reversal of character identities. However, most discussions focus on copyright and ethics, or are limited to specific instances, not deeply investigating the internal mechanism of aesthetic reconstruction. By taking AI remixes of well-known intellectual properties such as Empresses in the Palace and Dream of the Red Chamber as examples, this study applies Bakhtin's theory of carnival and Baudrillard's theory of simulacrum to explore how aesthetic reconstruction occurs when classic works are significantly changed by radical AI adaptation.