Research on the double-edged sword impact of AI performance feedback on employee creativity

Abstract

Based on Elaborated Information Processing Theory with AI-specific mechanisms, this study explores the “double-edged sword” effect of AI performance feedback on employees' creativity through two paths, namely the “positive effect” (via critical reflection) and the “negative effect” (via cognitive inertia), as well as the key moderating role of core self-evaluations (CSE). This study employed a rigorous single-source, three-wave survey design to investigate the phenomenon. The empirical investigation was conducted among high-tech enterprises in Shandong, Beijing and other regions, yielding 249 valid employee questionnaires after thorough data-cleaning procedures to eliminate invalid responses. This study revealed a dual-mediation mechanism in which AI performance feedback influenced employee creativity through both critical reflection and cognitive inertia. Importantly, core self-evaluations (CSE) served as a crucial moderator: high-CSE individuals exhibited stronger positive effects through enhanced critical reflection, while low-CSE individuals demonstrated more pronounced negative effects via increased cognitive inertia. Generative AI and traditional algorithmic AI systems showed differential impacts on the two pathways. The study addresses the theoretical gap in prior literature, which either overemphasises the benefits or overstates the harms of AI performance feedback without clarifying its dual cognitive mechanisms. It deepens theoretical and practical research on the impact of AI performance feedback on employees' creativity by distinguishing AI performance feedback from traditional feedback, unpacking heuristic versus systematic processing under algorithmic inputs, and introducing granular AI characteristics (frequency, quality, redundancy) into the dual-path model.