생성형 AI 건강정보 광고에서 인간 개입 정보가 광고 신뢰에 미치는 영향: 지각된 전문성의 매개효과를 중심으로

초록

As the use of generative AI in advertising production continues to expand, there is growing interest in how AI disclosure formats influence consumer responses. Particularly in health information advertising, where accuracy and credibility are critical, it is necessary to examine the effects of disclosure formats that explicitly indicate whether human experts have reviewed and revised the content. Accordingly, this study examined differences in consumers’ perceived expertise and advertising trust according to AI disclosure type (Simple AI Disclosure vs. Human Intervention Disclosure) and investigated the mediating effect of perceived expertise on the relationship between human intervention disclosure and advertising trust. A single-factor, between-subjects experiment was conducted with 100 consumers aged 50 and older who were randomly assigned to one of two experimental conditions. Data were analyzed using SPSS Statistics 29.0 and PROCESS Macro (Model 4), employing independent-samples t-tests and bootstrapping-based mediation analysis. The results showed that the Human Intervention Disclosure condition produced significantly higher levels of perceived expertise and advertising trust than the Simple AI Disclosure condition. In contrast, the indirect effect through perceived expertise was not statistically significant. These findings suggest that although human intervention disclosure enhances perceived expertise, perceived expertise alone may not sufficiently explain the process through which advertising trust is formed, indicating the need to consider additional factors related to trust formation. The findings also demonstrate that providing information about expert review and revision can be effective in enhancing consumers’ perceived expertise and advertising trust in generative AI-based health information advertising. This study contributes to both research and practice by empirically demonstrating the effects of disclosing human expert involvement beyond the mere disclosure of AI use through a between-subjects comparison.

키워드