research
Distrust Transfers Between People, Not Algorithms
A bad interaction lowered trust in the next unfamiliar human provider but not in the next AI agent.
Summary
A bad interaction lowered trust in the next unfamiliar human provider but not in the next AI agent.
University of Texas experiments found that participants generalized betrayal across people in similar service roles, while an unfair algorithm did not reduce trust in a later algorithm. The laboratory games isolate one social effect and do not measure every consumer or AI interaction.
Why it matters
A bad interaction lowered trust in the next unfamiliar human provider but not in the next AI agent.
Limits and context
- University of Texas experiments found that participants generalized betrayal across people in similar service roles, while an unfair algorithm did not reduce trust in a later algorithm.
- The laboratory games isolate one social effect and do not measure every consumer or AI interaction.
Key claims
A bad interaction lowered trust in the next unfamiliar human provider but not in the next AI agent.
Qualification: University of Texas experiments found that participants generalized betrayal across people in similar service roles, while an unfair algorithm did not reduce trust in a later algorithm.
Evidence: source-2026-07-21-021
Sources
- University of Texas at Austin via EurekAlert: Distrust transferUniversity of Texas at Austin via EurekAlert · official announcement
Corrections
No corrections have been recorded for this story.