Search engine versus affective thought-guided chatbot for knowledge acquisition: the mediating roles of perceived social, expressive and diagnostic transparency and cognitive effort
通过实验比较搜索引擎、简单聊天机器人、思维引导型和情感思维引导型聊天机器人对知识学习效果和持续使用意愿的影响，发现情感思维引导型聊天机器人学习效果最佳，并揭示了感知透明度与认知努力的中介机制。
Purpose This study aims to investigate the effects of different types of chatbots on knowledge learning performance and continuous usage intention by designing thought-guided and affective thought-guided chatbots. Drawing on coping theory, the study further seeks to uncover the underlying mechanism, focusing on perceptions of three nuanced aspects of transparency—social, expressive and diagnostic—and cognitive effort in human–artificial intelligence interactions. Design/methodology/approach Two experiments were conducted to investigate the impacts. A within-subjects experiment compared search engines and popular simple chatbots. A between-subjects experiment added thought-guided and affective thought-guided chatbots. Analysis of variance and partial least squares structural equation modeling were employed to analyze the differences and relationships. Findings Results indicate that, despite users’ initial preference for selecting chatbots, there are no differences in terms of knowledge learning performance and continuous usage intention between the search engine and the simple chatbot. Among the four tools for knowledge acquisition, the affective thought-guided chatbot achieves the best knowledge learning performance, and the thought-guided chatbot shows the lowest continuous usage intention. The mechanism test reveals that, compared to the search engine, the affective thought-guided chatbot increases users’ perceptions of all three aspects of transparency and cognitive effort. Whereas the thought-guided chatbot only increases perceived expressive transparency and cognitive effort. Furthermore, perceived expressive transparency and perceived diagnostic transparency improve knowledge learning performance, perceived social transparency and perceived diagnostic transparency promote continuous usage intention and cognitive effort enhances knowledge learning performance but weakens continuous usage intention. Originality/value This study contributes to the literature by concretizing coping theory in the chatbot context and dissecting interaction transparency into three nuanced aspects from a more detailed perspective. It addresses the ongoing debate over whether chatbots will replace search engines. It further introduces two innovative types of chatbots and presents a comprehensive framework to reveal the underlying mechanism.