Abstract: Exploring the complex logical mechanisms behind AIGC-driven deepfake information generation has significant practical value for constructing a cognitive framework for understanding deepfake information and formulating targeted governance strategies in cyberspace. Drawing on the actor-network theory, this study constructs a theoretical framework for analyzing AIGC deepfake information generation, focusing on four aspects: problem presentation, allocation of benefits, mobilization, and exclusion of dissent. It further interprets the dynamic process of deepfake information generation in terms of network formation, alliance-building, and stabilization. The findings indicate that the continuous output of AIGC-generated deepfake information is likely to intensify adverse social effects, such as technological domination, truth decay, and moral dissolution. The production and dissemination of deepfake information involve AIGC technologies translating heterogeneous actors through interest-driven strategies, driving the deepfake interest network from formation to stabilization while engaging in a competitive dynamic with opposing organizations. Based on these findings, this study proposes targeted governance strategies for AIGC deepfake information across four dimensions: moral governance, rule of law, technological governance, and crowd-based governance.

Key words: AIGC, Deepfake information, Generative mechanisms, Actor-network theory, Human-technology collaboration