Does Organizational Support Still Matter for Artificial Intelligence Adoption? Evidence from a Quasi-experimental Study in a Thai Commercial Bank

Abstract

Artificial intelligence (AI) is increasingly embedded in organizational workflows, and employees can access and experiment with AI tools through ‘Bring Your Own AI’ practices. Wider access raises a theoretical question about whether organizational support (OS) continues to guide technology acceptance once AI access is not institutionally constrained. Prior research has examined OS in Technology Acceptance Model 3 (TAM3) mainly through correlational and structural equation modelling-based designs, and causal inference remains limited. A quasi-experimental, posttest-only 2 × 2 vignette design was used in a Thai commercial bank with 591 employees to address this gap. Participants received either a supportive or a non-supportive organizational scenario, and scenarios were assigned across frontline and back-office roles. Attitudes towards AI and intentions to use AI were measured and analyzed using a two-way multivariate analysis of variance. Results indicate a significant causal effect of OS on both outcomes, and job type moderated the effects with stronger gains among back-office employees. The findings align with TAM3’s intervention logic and refine its application through job type as a task-based boundary condition in AI-enabled work settings. The study also strengthens causal inference in AI adoption research and clarifies how OS guides employee responses under widespread AI access.