代言人類型與 AI 敘述對綠色產品態度及廣告效果的影響：代言人可信度的中介作用

Abstract

隨著人工智慧 (Artificial Intelligence, AI) 技術快速發展，AI 已逐漸成為企業塑造產品價值與品牌形象的重要元素，虛擬代言人 (Virtual Endorser) 亦成為行銷中的新興趨勢。然而，虛擬代言人雖具科技感優勢，在情感信任建立上仍不及真人代言人。另一方面，綠色產品因價格較高且存在資訊不對稱，使消費者更加依賴外部資訊線索進行消費決策。因此，本研究欲探討 AI 敘述與代言人類型如何影響綠色產品之廣告效果與消費者購買決策。 本研究首先透過探索性訪談結合文獻建構研究架構與假說，建立兩個主要研究模型與三項補充分析。模型一聚焦於 AI 敘述如何透過認知與信任轉換機制影響綠色產品之廣告效果，並檢驗代言人類型之調節效果；模型二則探討代言人可信度如何進一步影響消費者購買意願，並檢驗環保意識之調節效果。此外，補充分析則進一步比較不同廣告情境下真人與虛擬代言人所造成之差異。 本研究採用 2（產品訊息類型：有 AI 敘述 vs. 無 AI 敘述）× 2（代言人類型：真人代言人 vs. 虛擬代言人）之實驗設計，以環保運動鞋作為廣告情境，共蒐集 213 份有效樣本。資料分析採用路徑分析、PROCESS 調節式中介分析、單純主要效果分析及雙因子變異數分析，以檢驗各項假說。 研究結果顯示，廣告中的 AI 敘述能有效提升消費者對產品 AI 科技含量之感知，進而提高產品診斷性與代言人可信度，並進一步影響購買意願、廣告態度及企業品牌形象。研究亦發現，代言人類型會影響上述效果之強弱，其中以虛擬代言人有更強的認知轉化效果。此外，消費者環保意識亦會影響綠色產品廣告的說服效果，其中高環保意識者會展現較強的購買轉化效果。後續補充分析亦發現，在未提供 AI敘述的情境下，虛擬代言人本身即可作為科技感的重要線索，顯著提升消費者對產品 AI 科技含量的感知。 然而，不論廣告中是否加入 AI 敘述，真人代言人在真誠度、可信度與專業度評價上皆顯著高於虛擬代言人，且能帶來較佳的企業品牌真誠度評價，顯示虛擬代言人雖具科技適配優勢，但在建立情感信任與品牌真誠形象方面，仍難以完全取代真人代言人。 綜合而言，本研究證實，綠色產品的行銷效果會受到訊息內容、代言人類型與受眾特質三者交互作用的共同影響。本研究補充了 AI 線索、虛擬代言人與綠色行銷領域之研究，首次提出及驗證AI敘述能提升產品診斷性，並闡明產品診斷性影響代言人可信度的作用機制，以及虛擬代言人與真人代言人在科技適配和情感信任上的不同優勢，亦為企業在 AI 時代下之綠色產品行銷策略提供代言人選擇、提升產品診斷性及重視高環保意識消費者等實務參考。

With the rapid development of Artificial Intelligence (AI) technology, AI has gradually become an important element for firms in shaping product value and brand image. At the same time, virtual endorsers have emerged as a growing trend in marketing. However, despite their technological appeal, virtual endorsers remain less effective than human endorsers in establishing emotional trust. Meanwhile, due to higher prices and information asymmetry associated with green products, consumers tend to rely more on external informational cues when making purchase decisions. Therefore, this study explores how AI-powered product features and endorser type influence advertising effectiveness and consumer purchase decisions for green products. This study first developed the research framework and hypotheses through exploratory interviews combined with a literature review, resulting in two primary research models and three supplementary analyses. Model 1 focuses on how AI narratives influence advertising effectiveness for green products through cognitive and trust transformation mechanisms, while examining the moderating effect of endorser type. Model 2 investigates how endorser trustworthiness further affects consumers’ purchase intention and examines the moderating role of environmental consciousness. In addition, supplementary analyses further compare the differences between human and virtual endorsers under different advertising scenarios. This study employed a 2 (advertising narrative type: with AI narrative vs. without AI narrative) × 2 (endorser type: human vs. virtual) experimental design, using an eco-friendly sneaker as the advertising scenario. A total of 213 valid samples were collected. Data were analyzed using path analysis, PROCESS moderated mediation analysis, simple main effect analysis, and two-way ANOVA to test the hypotheses. The results indicate that the inclusion of AI narratives in advertisements can effectively enhance consumers’ perceived level of AI technology in the product, which in turn improves product diagnosticity and endorser trustworthiness, thereby further influencing purchase intention, advertising attitude, and corporate brand image. The findings also show that endorser type affects the strength of these relationships. Specifically, virtual endorsers generate stronger cognitive transformation effects. Furthermore, consumers’ environmental consciousness also influences the persuasive effectiveness of green product advertisements. Consumers with higher environmental consciousness exhibit stronger purchase conversion effects. Supplementary analyses further reveal that, without explicit AI narratives, virtual endorsers themselves can serve as an important technological cue, significantly enhancing consumers’ perceived level of AI technology in the product. However, whether AI narratives are included in advertisements, human endorsers receive significantly higher evaluations in terms of sincerity, trustworthiness, and expertise than virtual endorsers. They also generate more favorable perceptions of corporate brand sincerity, suggesting that although virtual endorsers possess advantages in technological congruence, they still cannot fully replace human endorsers in building emotional trust and conveying a sincere brand image. Overall, this study confirms that the marketing effectiveness of green products is jointly influenced by the interaction among message content, endorser type, and audience characteristics. This study extends the literature on AI cues, virtual endorsers, and green marketing by being the first to propose and empirically validate that AI narratives enhance perceived diagnosticity. It further explains the mechanism through which perceived diagnosticity influences endorser trustworthiness and reveals the distinct advantages of virtual and human endorsers in terms of technological congruence and emotional trust. Moreover, the findings provide practical implications for firms in developing green marketing strategies in the AI era, including guidance on endorser selection, enhancing perceived diagnosticity, and addressing the needs of consumers with high environmental consciousness.