| 摘要: | 生成式AI聊天機器人是目前極具商業潛力的顛覆性創新，由於生成式聊天機器人的崛起，讓人們對於人機互動的方式和體驗有了新的期待和挑戰。因此目前生成式AI成為各界相當熱門的討論和研究議題，然而目前仍較缺乏生成式AI聊天機器人對使用者態度和看法的量化研究。
故本研究進行實證研究，探討生成式AI聊天機器人擬人化特性對使用者互動意願的影響。以力場分析理論為基礎，考察了生成式AI聊天機器人擬人化特徵（同理心、適應性、感知溫暖）等驅動力因素和（溝通延遲和感知冷淡）等制約力因素對感知溝通品質（準確性、可靠性、能力）、使用者好感度和人機互動意願的影響，以及隱私擔憂在此應用背景下的調節效果。研究方法分為兩部分，第一部分欲了解大眾生成式AI使用經驗，第二部分透過情境實驗，了解作為行銷和娛樂用途之聊天機器人對話風格對使用者態度的影響。兩部份皆透過問卷調查將有效樣本資料進行結構方程模型的分析。研究結果顯示同理心、適應性和感知溫暖對消費者對聊天機器人的感知溝通品質產生正面影響，而溝通延遲和感知冷淡則會產生負面影響。整體而言，若使用者感受聊天機器人有較高的溝通品質，將對使用者好感度產生正面影響，進而有更高的人機互動意願。儘管隱私問題在我們提出的模型中沒有發揮調節作用，但結果也揭示了某些程度的隱私悖論。
研究提供了對於生成式AI聊天機器人擬人化特性及其對使用者影響的新洞察，並為企業利用生成式AI聊天機器人作為品牌行銷與顧客服務提供一些見解和指導方向，為後續學者和開發者提供了理論和實務建議。
;Generative AI chatbot is a highly disruptive innovation with significant commercial potential. The rise of generative chatbots has introduced new expectations and challenges regarding human-computer interaction. Consequently, generative AI has become a popular topic for discussion and research. However, there is still a lack of quantitative research on user attitudes and perceptions toward generative AI chatbots. This study conducts empirical research to explore the impact of anthropomorphic design in generative AI chatbots on users′ willingness to interact. Based on Force Field Theory, the study examines the driving factors such as the anthropomorphic characteristics of generative AI chatbot (empathy, adaptability, perceived warmth) and the restraining factors (communication delay, perceived coldness) on perceived communication quality (accuracy, credibility, competence), user likability, and willingness to interact. It also investigates the moderating effect of privacy concerns in this context. The research methodology is divided into two parts. The first part aims to understand the public′s experience with generative AI, while the second part uses situational experiments to understand the impact of chatbot communication styles for marketing and entertainment purposes on user attitudes. Both parts employ questionnaire surveys, with the collected data analyzed using structural equation modeling. The results indicate that empathy, adaptability, and perceived warmth positively influence consumers′ perceived communication quality with chatbots, whereas communication delay and perceived coldness have negative effects. Overall, higher perceived communication quality leads to increased user likability and greater willingness to interact with the chatbot. Although privacy concerns did not moderate the proposed model, the findings revealed a certain degree of privacy paradox.
This study provides new insights into the anthropomorphic characteristics of generative AI chatbots and their effects on users. It offers practical guidance for businesses using generative AI chatbots for branding and customer service, and provides theoretical and practical recommendations for future researchers and developers. |
