Research on Users’ Trust in Customer Service Chatbots Based on Human-Computer Interaction
Abstract
As Human-Computer Interaction (HCI) moves towards deep collaboration, it is urgent to study users’ trust in chatbots. This study takes customer service chatbots as an example. Firstly, literature review is conducted on the relevant research on users’ trust in chatbots, and the value chain model of customer service chatbots is analyzed. Taking Taobaoxiaomi as the specific research object, we conducted in-depth interviews with 18 users, organized the interview data with value focused thinking method (VFT), constructed the users’ trust model of customer service chatbots, and carried out an empirical test by questionnaire survey. The results show that professionalism, response speed and predictability have positive effects on users’ trust in chatbots, while ease of use and human-likeness have no significant positive effects on users’ trust. Besides, brand trust has a positive impact on users’ trust in chatbots, risk perception negatively affects users’ trust in chatbots, and human support has no significant negative effect on users’ trust. Finally, privacy concerns have a moderating effect on environmental factors (brand trust, risk et al.). This study will deepen the understanding of human-computer trust and provide reference for the industry to improve chatbots and enhance users’ trust.
随着人机交互（HCI）走向深度合作，研究用户对聊天机器人的信任是非常迫切的。本研究以客户服务聊天机器人为例。首先，对用户对聊天机器人信任的相关研究进行了文献回顾，并分析了客服聊天机器人的价值链模型。以Taobaoxiaomi为具体研究对象，对18位用户进行了深度访谈，用价值聚焦思维法（VFT）整理访谈数据，构建了客服聊天机器人的用户信任模型，并通过问卷调查进行了实证检验。结果显示，专业性、响应速度和可预测性对用户对聊天机器人的信任有积极影响，而易用性和人性化对用户的信任没有明显的积极影响。此外，品牌信任对用户对聊天机器人的信任有积极影响，风险认知对用户对聊天机器人的信任有消极影响，而人类支持对用户的信任没有明显的消极影响。最后，隐私问题对环境因素（品牌信任、风险等）有调节作用。本研究将加深对人机信任的理解，为业界改进聊天机器人、提高用户信任度提供参考。