GenAI dependence and team innovation in construction projects: moderated mediation analysis of knowledge integration, team reflexivity and learning goal orientation

Purpose Generative artificial intelligence (GenAI) is increasingly embedded in construction project work, yet its implications for project team innovation remain insufficiently understood. Moving beyond a simple adoption or use perspective, this study examines how perceived team-referent GenAI dependence is associated with team innovation in construction projects. Design/methodology/approach Drawing on information processing theory, this study develops a moderated mediation model linking GenAI dependence, knowledge integration, team reflexivity, learning goal orientation and team innovation. Based on survey data from 257 construction project professionals in China, the model was tested using structural equation modeling, supplemented by hierarchical regression and bootstrapped analysis. Findings Based on survey data from the Chinese construction industry, the results show that GenAI dependence is positively associated with knowledge integration and negatively associated with team reflexivity. These two team processes account for opposing indirect associations between GenAI dependence and team innovation. Team learning goal orientation further attenuates the negative association involving team reflexivity, along with the corresponding negative indirect association with team innovation. Originality/value By examining perceived team-referent GenAI dependence as an information-processing condition rather than simply a higher degree of GenAI use, this study identifies the opposing pathways and boundary conditions associated with its enabling and constraining implications for innovation in construction project teams, contributing to a more comprehensive and timely understanding of AI implications in the construction industry.