Recombinant myth: the generative AI adaptation ecosystem of Ne Zha 2 on Douyin

This article examines the emergent practice of generative AI (GenAI) adaptation through a case study of short-video reinterpretations of the Chinese animated film Ne Zha 2 on Douyin. Drawing on comparative textual analysis and semi-structured interviews with key creators, I conceptualize this phenomenon as a GenAI adaptation ecosystem. I argue that this ecosystem is structured by the dialectical tension between two central axes. First, on the Creator Axis, I challenge the narrative of democratization, demonstrating how GenAI paradoxically establishes a new, skill-based hierarchy and reconfigures authorship into a precarious human–AI entanglement. Second, on the Platform Axis, I reveal how the algorithmic–AI confluence constitutes a regime of governance that drives systemic homogenization. Here, the pursuit of interpretive depth is subordinated to the imperative of algorithmic optimization, pressuring creators towards recombinant appropriation over originality. Ultimately, this ecosystem signals a recursive turn in cultural history, marking a return to an ‘Algorithmic Middle Ages’ which borrows from André Bazin’s vision of medieval art where authorship becomes a fluid, collective, yet profoundly precarious achievement. This yields a theoretical paradox for the concept of intertextuality characterized by an infinite expansion of its form alongside a drastic implosion of its meaning.