Establishing Antecedents and Outcomes of Human-AI Collaboration: Meta-Analysis

Human interactions with Artificial intelligence (AI) have gained significant attention in recent years. However, the extant literature fails to identify the relative importance of various antecedents and outcomes related to human-AI collaboration (HAIC). We used a meta-analysis approach using 248 effect sizes from 84 relevant papers related to HAIC. The study uses the Technology-Organization-Environment-Individual (TOEI) framework to categorize different antecedents related to HAIC. Additionally, we classified these antecedents into enablers and barriers to understand the impact on HAIC. For instance, trust in AI technology enables HAIC, whereas risk associated with AI use is a barrier to HAIC. Further, we analyzed the outcomes of HAIC at the organizational and individual levels. Subgroup analysis is carried out to examine the heterogeneity based on the type of economies. This study’s novelty lies in integrating the divergent research findings related to HAIC and using meta-analytic methods to generate multifaceted insights for practitioners and academia.