Source: Erengin et al. (2024) — “You, Me, and the AI: The Role of Third-Party Human Teammates for Trust Formation Toward AI Teammates,” Journal of Organizational Behavior
Trust in AI sounds like a one-on-one relationship.
Me and the tool. Me and the chatbot. Me and the system that keeps confidently suggesting things at suspicious speed.
But in real teams, trust is social.
This paper studies how third-party human teammates shape trust toward an AI teammate. In plain English: if I see someone else on my team treating the AI as useful, competent, and trustworthy, that changes how I see it too.
This feels very real.
Most people do not form opinions about workplace technology in isolation. They watch what the credible people around them do. If the respected teammate uses the AI carefully and gets good results, the system feels safer. If that teammate rolls their eyes every time the AI appears, the AI is already losing before it speaks.
This matters because companies often treat AI adoption like an individual training issue.
Teach the user. Give the user a guide. Send the user a webinar link that they will absolutely not watch.
But if AI is entering team workflows, the social layer matters. People learn not only from documentation, but from other people’s behaviour. Trust travels through the team.
That is both good and dangerous.
Good, because trusted colleagues can help normalise careful use. Dangerous, because one influential sceptic can quietly poison adoption, sometimes for valid reasons, sometimes because the first demo was terrible and nobody recovered emotionally.
My takeaway: AI teammates need social onboarding, not just technical onboarding. Leaders should pay attention to who models the behaviour, who validates the tool, and who helps others interpret it. Trust is not installed. It spreads.