Key points are not available for this paper at this time.
As artificial intelligence (AI) increasingly permeates firms’ digital innovation, trust in AI has emerged as a core determinant of the effectiveness of human-AI collaborative innovation. However, prior research has not reached a consensus on the underlying mechanisms or boundary conditions of this relationship. Drawing on information processing theory (IPT), this study empirically examines the nonlinear relationship between trust in AI and digital innovation using survey data from 269 AI-savvy senior innovation managers. The results indicate an inverted U-shaped association, driven by the net effect of the “technology-driven” benefit mechanism and the “wicked curse” cost mechanism. Moreover, intellectual capital weakens this inverted U-shaped relationship, task complexity strengthens it, and data quality shifts the turning point to the right. Overall, these findings clarify the logic of under-trust and over-trust in AI and delineate contextual boundaries for interpreting its double-edged effect.
Lin et al. (Mon,) studied this question.