19698: The Effectiveness and Boundary Conditions of AI Adoption in Enterprise: A Meta-Analysis Study (session 25940)

Due to the widespread application of artificial intelligence (AI) by enterprises, examining the effect of AI technologies on enterprise growth has become a prominent topic in both academic and industry area. However, existing studies on the effect of AI adoption on innovation and performance at both the employee and enterprise levels have not reached consistent conclusions. This not only highlights the complexity of the effectiveness of AI adoption, but also reflects the limitations of current research in explaining its mechanisms and systematically outlining its boundary conditions. To address this, we employ a meta-analysis approach to systematically integrate 77 empirical studies, encompassing 88 independent investigations with a total sample size of 5,166,848, to comprehensively examine the effect of AI adoption on employee and enterprise innovation and performance, as well as the moderating factors involved. The results reveal that AI adoption exhibits a moderately strong positive correlation with enterprise innovation and employee innovation, while demonstrating a weak positive association with enterprise performance and employee performance. Moreover, factors such as the type of AI technology, employee roles, human-AI interaction methods, policy support, industry type, data collection time point, national computing power levels, and the measurement of AI adoption significantly moderate these relationships. By summarizing the universal conclusions regarding the effect of AI adoption on enterprises and unveiling the sources of variability, we provide a richer perspective and a more comprehensive explanatory framework for studying AI adoption, and offers a reliable theoretical basis for managers to make decisions on AI adoption.

view paper (if available)