Same Gossip, Different Judgment: Why Inappropriate Workplace Gossip Is More Acceptable When It Targets AI

Artificial intelligence is rapidly becoming an active member of organizational teams. Yet successful human–AI symbiosis depends on more than intelligent algorithms, trust calibration, or effective task allocation. As AI becomes a social interaction partner, organizations must also establish shared moral norms governing how artificial teammates are treated. Surprisingly, little is known about whether employees apply the same ethical standards to AI coworkers as they do to human colleagues or why these judgments differ. To address this question, we conducted two preregistered online vignette experiments (N = 546) in which participants observed coworkers engaging in identical inappropriate workplace gossip directed either at a human colleague or an AI teammate. Building on Bandura's theory of moral disengagement, we tested whether cognitive moral disengagement explains differences in observers' moral evaluations and willingness to intervene. Across both studies, participants consistently judged gossip about an AI teammate as more morally acceptable and reported significantly lower willingness to intervene than when the same behavior targeted a human colleague. AI also increased endorsement of all eight mechanisms of moral disengagement. Parallel mediation analyses showed that moral disengagement fully accounted for the effects of target identity on both outcome variables, indicating that AI changes the cognitive processes through which observers evaluate unethical workplace behavior. Although the specific mechanisms contributing unique indirect effects varied somewhat across studies, dehumanization and distortion of consequences emerged as the most robust mediators. An exploratory analysis further demonstrated that these effects generalized across male and female perpetrators. These findings extend research on human–AI symbiosis by identifying shared moral norms as a previously overlooked prerequisite for successful collaboration. They suggest that organizations should focus not only on designing trustworthy AI systems but also on fostering ethical workplace cultures in which inappropriate behavior is evaluated according to the behavior itself rather than whether its target is human or artificial.