How Do Social Bots Become the "Other" ? —Identity Camouflage of Generative AI in Social Platforms

Under the research paradigm of "data-driven, behavior modeling, mechanism explanation," the deep identity camouflage achieved by social bots through gen-erative AI technology has become a core risk intervening in the information ecology of social media. This paper takes 206 labeled social bot accounts on the Weibo platform as the research object, integrates 4823 valid post data points and multimodal visualization results, and constructs a quantitative analysis framework from four dimensions: behavioral consistency, content homogenei-ty, metadata contradiction, and camouflage depth. The study finds that Weibo social bots achieve identity camouflage through four major strategies: "dynamic temporal mimicry," "entertainment content cloning," "cluster interaction fab-rication," and "superficial persona construction." Furthermore, generative AI technology drives their camouflage from "mechanical replication" to "semanti-cally coherent, lifelike narration." The research not only provides a three-dimensional quantitative indicator of "behavior-content-metadata" for social bot detection on Chinese platforms but also supplements the theoretical gap of "machine communicator identity construction" in computational com-munication studies, offering empirical support for platform governance and AI content regulation.