聊天機器人性別形塑對使用者信任感受之影響

Gendering chatbots by design and its impacts on users trust towards shopping chatbots

隨人工智慧(AI)科技的成熟，企業紛紛應用聊天機器人(chatbot)於其行銷服務當中。而大量AI服務的出現，讓人機互動領域學者開始重視AI的性別化議題，希望理解形塑AI性別的設計手法與AI性別化對使用者的影響。本研究透過兩階段實驗，分別探索不同聊天機器人的視覺設計作為性別提示的效果，以及購物情境中聊天機器人性別對於使用者信任感受之影響。前導實驗先檢視聊天機器人頭像設計中不同視覺元素對使用者性別感受之影響。再進一步藉由操作聊天機器人的性別提示組成四種條件的組間設計實驗，招募受測者進行模擬網路購物的互動實驗，了解不同性別購物聊天機器人給人的信任感受差異。結果發現聊天機器人頭像的髮長、顏色、與眉型都可以明顯影響人們對聊天機器人的性別感受；但聊天機器人的性別對於人們的信任感受僅有些微影響；反倒是受測者本身的性別與年齡對聊天機器人的信任感受影響較大。

With the rise of artificial intelligence technology, companies have integrated chatbots into e-commerce marketing services nowadays. With the emergence of avast number of AI applications, the researchers in HRI have begun to pay attention to the gender issue of AI, hoping to understand the impact of AI gender on users. This study investigates the effects of chatbot gender on user trust in online shopping contexts using atwo-stage experimental design. First, this study examined the impact of the visual elements of AI avatars on the user's gender perception in the pilot study. Then, this study conducted abetween-subject design experiment to examine how gender factors affect user trust towards shopping chatbots. The results found that gender cues have apparent impacts towards them; instead, the gender and age of participants have more impact on the user trust.