Mustafa Akben, Su Dong
The rapid emergence of generative AI is transforming how employees engage with technology to perform tasks, make decisions, and create value. Despite its transformative potential, empirical findings on AI adoption remain inconsistent, particularly regarding how job characteristics shape employees' confidence and readiness to use generative AI. Grounded in the Task-Technology Fit framework and self-efficacy theory, this research examines the curvilinear relationship between job complexity and AI self-efficacy and its subsequent effects on AI adoption readiness and behavior. We conducted two survey studies to test the proposed hypotheses using structural equation modeling. Results reveal that employees in both low- and high-complexity roles exhibit a low level of AI self-efficacy and a subsequent lower level of AI adoption behaviors compared to those in moderately complex roles. These findings challenge the assumption that highly skilled roles typically lead AI integration and instead highlight the importance of aligning task structure with AI capabilities. This study advances theory by introducing a non-linear boundary condition to technology adoption and offers practical guidance for organizations to design jobs and training programs that cultivate confidence and foster sustainable human-AI collaboration.