[Purpose/significance] In the context of generative AI, Online Knowledge Communities (OKC) have intro⁃duced a large number of bots embedded with AI algorithms, posing new challenges for developing human-bot collabora⁃tion patterns to better improve the performance of knowledge social construction. [Method/process] Collect humanbot collaboration network log data from 136 human users and 48 bot users in Wikipedia, and use grounded theory for multi case analysis. [Result/conclusion] The study establishes a human-bot collaborative knowledge social construc⁃tion model for OKC in generative AI contexts, revealing the mechanism of human-bot collaborative knowledge social construction, elaborating the effect of three basic task dependencies(flow,fit,sharing) and human-bot team resources re⁃quired for task execution on the formation of two human-bot collaboration patterns, automated assistance and enhanced assistance, as well as the important impact of these two human-bot collaboration patterns on the knowledge social con⁃
struction performance.From the perspective of human-bot collaboration, the research enriched and developed the coor⁃dination theory and social constructivist theory, providing useful guidance for human collaborating strategies and knowl⁃edge social construction performance improvement strategies in generative AI contexts.