Trust in AI vs human expert : investigating behavioral intention among employees

Abstract

Purpose: This study aims to investigate the interactive effect between information quality (high vs low) and agent (artificial intelligence [AI] expert vs human expert) on hospitality employees’ trust and behavioral intention. Design/methodology/approach: Five scenario-based survey experiments were conducted, systematically varying the combinations of information quality and agent. Data were collected from hospitality employees to examine their responses regarding trust, perceived effectiveness and behavioral intention. Findings: Low-quality information is more likely to be trusted and followed when it comes from an AI expert rather than a human expert. However, critical thinking skills and cognition-based trust may help mitigate this unfavorable effect. The findings demonstrate that both the information quality and the agent delivering it are critical determinants of employees’ psychological state and behavior in workplace settings where AI is integrated. Originality/value: This study contributes to automation bias and source credibility theory in hospitality by investigating the interplay between information quality and agent. It enriches the literature on AI’s integration in workplace settings and provides actionable insights for managers seeking to ethically implement AI technologies.