This, That & The Other: A Study in When We Trust Algorithms vs. Human Perception

Why Read This from an HCI Perspective

This piece is meaningful for HCI researchers and UX practitioners because it explores when and what people trust more among algorithms, intuition, or neither. It may look like a simple preference survey, but it actually touches on core themes such as decision delegation, trust formation, explainability, and automation bias. It’s a useful starting point when thinking about designing for the adoption of AI-based services or designing for users’ sense of control.

CIT's Commentary

From CIT’s perspective, what matters here is not the binary question of whether people ‘trust algorithms,’ but rather asking ‘what judgment resources are delegated in which situations.’ Human intuition is not just irrationality; it’s the result of accumulated contextual experience. Likewise, trust in algorithms depends not only on performance, but also on transparency, the cost of errors, and where responsibility lies. Therefore, rather than lightly categorizing user attitudes, such surveys should treat as design variables the uncertainty of the task, the level of risk, and the degree of explanation required. Practically, before setting the default for AI recommendations, it’s necessary to consider how control is exposed to users and in which cases human intervention should be placed front and center.

Questions to Consider While Reading

• Q.Are the scenarios presented in this survey designed so that task risk, uncertainty, and personal stakes are sufficiently distinguished?
• Q.How do you plan to tell whether a respondent’s ‘algorithm preference’ reflects actual trust—or instead convenience or reduced cognitive burden?
• Q.When translating the results into real UX design, could you also derive under what conditions human intervention or explanation is needed?