Opening the door to a holiday rental after a long journey only to find it hasn’t been cleaned as promised is enough to leave anyone feeling betrayed. According to new research led by Annabelle Roberts, assistant professor of marketing at the McCombs School of Business at The University of Texas at Austin, that experience can have lasting consequences. Rather than simply distrusting the property owner responsible, consumers are more likely to approach their next holiday rental host with suspicion. Similar reactions can occur after receiving incorrect information from a customer service representative or being let down by a rideshare driver. “We’re trusting people all the time, whether we think about it that way or not,” Roberts says. “There are many situations where consumers need to decide whether to trust an unknown person, and they may use previous interactions with other people in similar settings as a reference point.”

To understand how trust transfers from one interaction to the next, Roberts and colleagues Emma Levine and Jane Risen of the University of Chicago conducted 21 studies involving nearly 12,000 participants. Ten studies used online trust games in which participants exchanged money with one partner before making trust decisions with another. The remaining studies presented everyday situations, such as asking a co-worker to keep a secret or lending an item to a neighbour. Participants then rated how willing they would be to trust a different person in a similar situation after learning whether their trust had been honoured or betrayed.

Across all 21 studies, the researchers found that even a single interaction shaped future expectations. Positive experiences increased participants’ willingness to trust someone else in a comparable setting, while negative experiences reduced it. Rather than judging each new person independently, people carried lessons from previous encounters into future interactions. The findings suggest that trust is not built from isolated events but from an accumulation of social experiences that influence expectations of others.

Roberts identified two reasons for this pattern. The first is emotional: being deceived or let down is unpleasant, making people more cautious about exposing themselves to similar risks in the future. The second is perceptual. People naturally categorise others into groups, meaning that a bad experience with one rideshare driver, holiday rental host, or service representative can shape expectations of others in the same role. “You’re updating your beliefs about the population with each interaction that you have,” Roberts explains.

The studies also showed that distrust develops more readily than trust. A negative interaction had roughly twice the impact of a positive one, particularly when someone deliberately exploited another person’s trust. Although the research focused on trust between individuals rather than organisations, the findings have important implications for businesses. Companies and service providers may encounter customers whose scepticism stems not from their own actions but from previous negative experiences with similar businesses. Responding promptly to problems and demonstrating reliability can help rebuild confidence over time.

Interestingly, the same pattern did not apply to artificial intelligence. When participants interacted with algorithms in trust games, even unfair outcomes did not reduce their willingness to trust other AI systems. Roberts suggests this is because people do not view interactions with AI as social in the same way they do interactions with humans. “This trust effect is specific to social interactions,” she says. “People don’t lump AI agents together, because they don’t see those interactions as social.” The findings highlight how strongly human relationships shape consumer trust—and how difficult it can be to overcome the effects of a single negative experience.

More information: Annabelle Roberts et al, Learning to distrust: One trust experience changes the expected value of trust, Journal of Experimental Social Psychology. DOI: 10.1016/j.jesp.2026.104930

Journal information: Journal of Experimental Social Psychology Provided by University of Texas at Austin