From bad omens to commodities: Uncovering the spatiotemporal patterns and socio‐economic factors of online crow trade on Baidu Tieba using artificial intelligence

Abstract High demand for wild birds as pets stimulates overexploitation and threatens global avian biodiversity. Unlike commonly traded birds valued for colourful plumage or attractive songs, crows (e.g. Corvus , Pyrrhocorax and Coloeus ) have increasingly appeared in online records of rearing, acquisition and trade, yet these activities remain overlooked. To investigate these activities using massive online datasets, we developed a large language model‐based pipeline with manual verification to identify the validity of 21,846 forum posts from Baidu Tieba, achieving an accuracy of 89.51%. The main motivations of crow rearing were flight training and companionship, relying on the high cognitive capabilities of crows rather than the aesthetic or vocal traits valued in traditional bird markets. Local pickup was the dominant transaction mode, with hotspots concentrated in prefectures with high per capita GDP, including Beijing, Shanghai, Chengdu and Xiamen. Corvus and Pyrrhocorax accounted for 14,426 recorded individuals and exhibited dense intercity trafficking networks. Routes followed a clear directional pattern, from origins with high wild crow sightings to economically developed southeastern China. The trafficking network of Coloeus may be at an early stage, but related records increased rapidly. This study provides evidence and data to support targeted law enforcement efforts to curb such potentially illegal activities. With adapted search phrases and prompts and manual validation, our pipeline could be extended to other species threatened by illegal online trade, integrating digital monitoring into routine conservation efforts. Read the free Plain Language Summary for this article on the Journal blog.