Stability through human perception: Technology acceptance models’ robustness across various interaction perspectives and comparable technologies

Technology acceptance research has traditionally focused on system design and interaction experience integrated into established frameworks like the Technology Acceptance Model (TAM; Davis, 1985 ) and the Unified Theory of Acceptance and Use of Technology (UTAUT; Venkatesh et al., 2003). While these models are widely applied to investigate the interaction between humans and technology, critiques have highlighted limitations concerning their generalizability, robustness, and conceptual clarity. This study addresses these challenges by examining the structural strength of TAM and UTAUT across comparable complex technologies and user interaction perspectives: between an active and a passive vignette. Furthermore, it introduces a model integrating Usage Perception (UP), incorporating the broader user perception of system usage, as a possible solution to the emerging vulnerability of traditional technology acceptance frameworks. A controlled laboratory experiment was conducted in a simulated online environment, featuring two functionally developed AI agents that varied in interaction type and user autonomy. This design allowed direct comparison between active and passive interaction user perspectives and comparable complex technologies across a broad and diverse sample of 311 participants. The results confirm that while some core predictors remained robust, traditional TAM and UTAUT structures showed inconsistency when disaggregated by interaction perspective and technology use case. In contrast, UP-integrated modelling showed greater path stability, theoretical coherence, and consistent predictive power. This paper offers a refined framework to holistically explain technology acceptance in complex AI systems—by integrating the human perception of system usage.