Trust and AI weight: human-AI collaboration in organizational management decision-making
Citation. Wen, Y., Wang, J., & Chen, X. (2025). Trust and AI weight: human-AI collaboration in organizational management decision-making. Frontiers in Organizational Psychology, 3, 1419403. https://doi.org/10.3389/forgp.2025.1419403
Summary. Across two scenario-based surveys of Chinese MBA students and managers, the authors examine how trust in AI relates to the percentage of managerial decision weight participants allocate to an AI agent. Willingness to collaborate mediates the trust–AI-weight relationship, while perceived AI free will weakens that indirect pathway in the performance-evaluation scenario.
Key claims
- In Study 1 (n = 111), higher trust in AI was associated with greater willingness to collaborate and greater AI decision weight in a simulated employee-recruitment decision; willingness to collaborate mediated the trust–weight relationship.
- In Study 2 (n = 210), conducted with a simulated employee-performance-evaluation decision, the direct effect of trust on AI weight was 7.72 percentage points (95% CI [4.06, 11.39]) and the conditional indirect effect through willingness to collaborate was smaller when perceived AI free will was high (3.33, 95% CI [1.30, 5.49]) than low (4.83, 95% CI [1.85, 7.74]).
- The moderated-mediation index for perceived AI free will was -0.67 (95% CI [-1.40, -0.24]), indicating that greater perceived AI autonomy weakened the association between trust and willingness to collaborate.
Evidence & limitations
- Peer-reviewed open-access journal article using two cross-sectional, scenario-based surveys of MBA students in southern China; participants included both managers and non-manager employees, so the evidence concerns stated allocation preferences rather than observed workplace decision rights or performance outcomes.
- The AI-weight outcome is a single 0–100% slider, and the authors note that the brief trust and free-will measures may not capture the constructs fully.
- The source supports a psychological mechanism in hypothetical recruitment and performance-evaluation tasks; it does not establish how organizations should calibrate, govern, or validate AI decision authority in deployed high-stakes HR systems.