Navigating the double-edged sword: an integrated TAM–SET model for sustainable knowledge sharing in the AI era

Purpose This paper aims to address the paradox where artificial intelligence (AI) enhances knowledge sharing efficiency but may erode user motivation. It develops and tests an integrated model synthesizing the technology acceptance model (TAM) and social exchange theory (SET) to explain the context-dependent drivers of sustainable knowledge sharing in the AI era. Design/methodology/approach A survey-based empirical design is employed. Data from 201 respondents were collected and analyzed using path analysis across three AI-related knowledge-sharing scenarios to examine the relationships among perceived usefulness, perceived ease of use, perceived benefits, perceived costs, sharing intention and knowledge-sharing behavior. Findings While perceived benefits and ease of use are stable drivers of sharing intention, the role of perceived cost is dynamic. It acts as a deterrent in direct human-AI interaction but transforms into a ‘sunk cost investment’ that promotes sharing in interpersonal contexts. A significant “intention-behavior” gap is also identified, stemming from poor technological affordances. Practical implications Actionable guidance is provided for platform designers and managers. Designers should create low-friction contribution channels and leverage “sunk cost” psychology. Managers should adopt context-specific incentive strategies and prioritize system usability to foster a sustainable knowledge ecosystem. Originality/value To the best of the authors’ knowledge, this is one of the first studies to empirically model the sustainability paradox of AI-driven knowledge sharing. It advances theory by demonstrating the context-dependent role of core constructs like “perceived cost,” extending TAM and SET to the human–AI collaboration domain.