Author(s)

Jiayang Sun, Qingya Liu, Yihan Li

Corresponding Author

Jiayang Sun

Abstract

This study investigates the efficacy of generative AI in organizational human-machine collaboration contexts, addressing significant theoretical gaps in understanding the interactive effects of task types and personal characteristics on collaborative efficiency. Integrating the Unified Theory of Acceptance and Use of Technology (UTAUT) and AI Self-Efficacy (AISE) frameworks, we develop a multidimensional theoretical model to elucidate the causal mechanisms and boundary conditions governing generative AI's impact on task performance. Employing a mixed experimental design in a simulated recruitment context (N=119), we conducted multilevel path analysis and moderation-mediation testing via SmartPLS and SPSS. Results reveal three key findings: (1) generative AI exhibits significantly stronger performance enhancement effects in structured versus unstructured tasks; (2) performance expectancy positively moderates the relationship between AI utilization and Human-AI capability development; and (3) Human-AI capability functions as a mediating mechanism between AI utilization and task performance. These findings contribute to the theoretical understanding of human-machine collaboration dynamics and provide empirical support for organizational AI deployment strategies that consider both task characteristics and individual user attributes.

Keywords

Generative AI, Task Performance, Performance Expectancy, Task Type, Human-AI Capability