Answers are grounded in this paper and cite their sources.
As artificial intelligence (AI) becomes an increasingly common source of advice, an important question is whether people respond differently to advice simply because it is identified as coming from AI. This study examined how explicitly labeling advice as AI-generated influences choice revision and decision confidence in everyday moral dilemmas. Participants made initial decisions and rated their confidence before receiving advice that either conflicted with or supported their choices. They then reconsidered their decisions and reported their confidence again. Participants were assigned to either an AI-labeled advice condition or a condition in which the advice source was unspecified, while the advice content remained identical. When advice conflicted with initial choices, participants were less likely to adopt AI-labeled advice than unlabeled advice, although the difference was not statistically significant. When advice agreed with initial choices, confidence increased regardless of the source label. In contrast, rejecting conflicting advice was associated with little change in confidence. Exploratory analyses also suggested different patterns of confidence change following advice adoption across source conditions, but these differences were not robust after statistical adjustment. Overall, the findings suggest that AI source labels may influence responses to conflicting advice, while highlighting the distinction between accepting advice and becoming more confident in one's judgment. Further research is needed to establish whether these patterns reflect algorithm aversion or other psychological processes.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: