How do consumers evaluate fashion information from artificial intelligence versus a human expert, and under what conditions is one more persuasive? This research examines evaluations of fashion information based on its source – human or AI – and how these evaluations, reflected in attitude and electronic word-of-mouth (e-WOM) (Study 1), are moderated by information type (Study 2; creative vs. analytical) and temporal focus (Study 3; past vs. future).
Three between-subjects experiments manipulated the source, information type and temporal focus of fashion information. Respondents were US residents aged 18 or older. Analyses were conducted via analysis of covariance (ANCOVA) and moderated mediation models.
Across studies, consumers consistently preferred human sources. Study 1 showed that human-generated fashion design trend information produced significantly more favourable attitudes and e-WOM. Study 2 found the human preference was significant in creative domains but not in analytical ones, consistent with humans' strength in aesthetic judgement. Study 3 demonstrated that the human advantage was significant in past-focused but not in future-focused content, underscoring humans' strength in interpreting past events.
Fashion retailers should align source types with content characteristics, emphasising human expertise in creative and retrospective contexts.
Integrating source credibility, algorithm aversion, mind perception and certainty–uncertainty framework, this research develops a unified explanation of when source effectiveness varies in fashion contexts and identifies boundary conditions where content characteristics (information type and temporal focus) align with the source to shape persuasive outcomes.
