생성형 AI 추천 광고 효과에 대한 추정된 미디어 영향력 가설의 검증 = Testing the Presumed Media Influence Hypothesis in Generative AI Recommendation Advertising

다국어 초록 (Multilingual Abstract)

This study challenges the limitations of conventional approaches that explain advertising effects in the context of generative AI–driven recommendation advertising solely through individual information processing or attitudinal evaluation. Because generative AI recommendation advertising delivers personalized messages while simultaneously being perceived as widely disseminated to others through the same algorithms and platforms, its persuasive effects are likely to be shaped not only by individual-level persuasion but also by social inferences regarding the influence of the advertising on others.
Accordingly, this study adopts Presumed Media Influence (PMI) theory as its core explanatory framework to empirically examine the effects of exposure to generative AI recommendation advertising, perceived personalization, and algorithm preference on the formation of PMI, as well as the structural pathways through which PMI leads to advertising attitudes and purchase intention.
Data collected through an online survey were analyzed using confirmatory factor analysis and structural equation modeling. The results demonstrate that the persuasive effects of generative AI recommendation advertising are formed not direct outcomes of individual information processing; rather, they are formed through the mediating role of presumed social influence. By applying PMI theory to the context of generative AI–based commercial advertising, this study extends the theoretical scope of advertising effects research and offers practical implications for the development of algorithmbased advertising strategies.