In the context of the digital economy, artificial intelligence (AI) is deeply embedded in organizational processes, making human–AI collaboration an emerging issue of growing interest among both scholars and practitioners. However, existing research remains fragmented and lacks systematic integration. To address this gap, this paper analyzes 220 English-language articles on human–AI collaboration. Using the BERTopic topic model, it identifies 13 research topics and, in combination with manual coding and hierarchical clustering visualizations, synthesizes the literature into four major domains: conceptualization and components, antecedents, outcomes, and boundary conditions. Building on this classification, this paper further conducts a systematic review of both English- and Chinese-language studies to develop an integrative research framework for human–AI collaboration. On this basis, it outlines future research directions, including deepening the understanding of human–AI collaboration and elucidating the dynamic mechanisms among its components, extending research on antecedents and revealing the joint effects of multiple elements, broadening outcome dimensions and examining cross-level linkages among outcomes, advancing research on boundary conditions, and enriching data sources and research methods. By systematically organizing and integrating the existing literature, this paper aims to provide a solid knowledge base and useful reference for future research on human–AI collaboration.