Pyramid Handoff: The Social Construction of Biopolitical Humanness of AI Humans in China

With the advancement of generative AI, AI humans are becoming increasingly human-like. Existing research assumes that humanness of AI humans is either “born” or freely crafted by adopters to satisfy audience preferences. However, these assumptions overlook the fact that, in China, humanness of AI humans is tightly regulated by both the state and platforms. Thus, this study proposes the theory of pyramid handoff and adopts an analytical metaphor of biopolitics to explain why different human-like traits of AI humans are governed by the state, platforms, or adopters, and how this multi-layered governance unfolds across the life stages of AI humans. Drawing on 23 months of fieldwork, this study demonstrates that humanness is governed as a biopolitical object across AI humans life course: selective breeding at birth, life optimization during function, and sovereign decision at death. Adopters can craft humanness only within the boundaries set by government policies and platform regulations. As a result, humanness of AI humans emerges as an uneven assemblage of state governance, platform-economic logics, and adopters’ usage objectives. This study provides the first systematic account of how humanness is socially constructed under a state–platform nexus in China.