聊天机器人是一种能够通过自然语言与用户进行对话的产品，能够快速响应用户提出的问题。随着人工智能技术的进步，聊天机器人在教育与学习领域展现出巨大潜力。在ChatGPT此类生成式人工智能聊天机器人产品出现后，目前的研究并未充分关注用户使用生成式人工智能聊天机器人产品的意愿和使用体验，特别是在学习场景下。本研究旨在通过探究用户在学习场景下对聊天机器人的使用意愿的影响因素，结合目前用户在学习场景下使用聊天机器人产品的需求以及痛点，将影响使用意愿的因素转化为设计策略，进行聊天机器人的产品设计，以提升用户的使用意愿以及使用体验。

    本研究分为四个部分。在第一部分，本研究基于技术接受模型(TAM)，从用户个人特征、产品体验以及用户与产品之间的关系三方面对技术接受模型进行扩展。将个人创新性、感知拟人化、社会临场感和感知信任四个影响因素添加至模型中，并提出了相应的假设。通过问卷调查的方式对用户在学习场景下聊天机器人的使用情况以及影响因素的数据进行收集，并运用偏最小二乘法结构方程模型对假设进行验证。

    在第二部分，本研究结合第一部分影响使用意愿的因素的假设检验结果，通过对用户进行一对一访谈，了解用户使用情况以及使用意愿影响因素在产品使用中的具体体现。并根据访谈结果提炼出用户在学习场景下使用聊天机器人的需求和痛点，通过用户旅程图进行梳理。

    第三部分，本研究根据前两部分的研究结果，提出了本研究的设计策略，包括交互过程清晰透明、降低提问难度、以及对话信息便捷管理，并根据设计策略产出具体的产品设计方案，设计相应功能对产品进行改进。

    在最后一部分，本研究进行了可用性测试实验。在实验中将本研究设计的聊天机器人产品与ChatGPT产品原型进行对比，通过对用户的使用意愿以及产品可用性评分数据进行分析，证明本研究提出的设计策略以及产品设计方案的有效性。

    本研究为国内探究用户对生成式聊天机器人使用意愿方面的研究提供了补充，在研究中提出的设计策略为基于生成式人工智能的产品设计方面提供了参考。

    Chatbots are the product that is capable of conversing with a user through natural language and responding quickly to questions posed by the user. With the advancement of AI technology, chatbots show great potential in the field of education and learning. After the emergence of generative AI chatbot products like ChatGPT, current research has not paid sufficient attention to users' willingness to use generative AI chatbot products and user experience of using these products, especially in learning scenarios. The purpose of this study is to explore the factors affecting users' intention to use chatbots in learning scenarios, combine the current users' needs and pain points of using chatbot products in learning scenarios, and translate the factors affecting intention to use into design strategies for chatbot product design in order to improve users' intention to use as well as the experience of using them.

    This study is divided into four parts. In the first part, based on the Technology Acceptance Model (TAM), this study extends the TAM from three aspects: personal characteristics of users, product experience, and the relationship between users and products. Four influencing factors, namely personal innovativeness, perceived anthropomorphism, social presence and perceived trust, are added to the model, and corresponding hypotheses are proposed. Data on users' use of chatbots in learning scenarios and the influencing factors were collected by means of questionnaires, and the hypotheses were verified using partial least squares structural equation modeling.

    In the second part, this study combines the hypothesis testing results from the first part regarding factors influencing intention to use. Through one-on-one interviews with users, the aim is to understand both their usage patterns and the specific manifestations of factors influencing intention to use during the use of product. Subsequently, users' needs and pain points in utilizing chatbots within learning scenarios are distilled based on the interview results, and organized using a user journey map.

    In the third part, this study proposes a design strategy for this study based on the results of the first two parts, including a clear and transparent interaction process, reducing the difficulty of asking questions, and convenient management of conversation information, and outputs a specific product design solution based on the design strategy, and designs the corresponding functions to improve the product.

    In the last part, usability testing experiments were conducted. In the experiment, the chatbot designed in this study is compared with the ChatGPT prototype, and the data on users' willingness to use the chatbot and the usability score of the chatbot are analyzed to prove the effectiveness of the design strategy and the product design solution proposed in this study.

    This study adds to the domestic research on users' willingness to use generative chatbots, and the design strategy proposed in this study provides a reference for product design based on generative AI.