AI의 개입 수준과 사회적 증거의 상호작용이 AI 기반 추천 콘텐츠의 수용도에 미치는 영향: 인지적 부담과 처리유창성의 순차적 매개효과

초록

Companies are actively adopting artificial intelligence (AI)-based recommendation systems to enhance consumer experiences. However, as the level of AI involvement increases, consumer resistance may also grow, leading to confusion about how to balance AI and human participation. This confusion often arises from the uncertainty inherent in AI technology, and consumers tend to rely more on external cues such as social proof rather than solely on information provided by AI. This study empirically examines how the interaction between AI involvement and social proof affects acceptance of AI-recommendations, as well as the underlying mechanisms. The results show that in the absence of social proof, acceptance is higher for recommendations generated through human-AI collaboration than for those generated by AI alone, while no significant difference is observed when social proof is present. Furthermore, this effect is sequentially mediated by cognitive load and processing fluency. This study demonstrates that the impact of social proof in AI recommendation environments varies depending on the degree of human involvement. It also clearly identifies, through consumers’ cognitive processes, when external cues are effective and when their influence is limited. Furthermore, the findings provide strategic guidance for effectively utilizing human involvement and social proof in practical service environments to simultaneously enhance organizational efficiency and consumer acceptance.

키워드