Research on User Interface of Conversational AI for Foreign Language Speaking

Objective: This study proposes a user interface for designing a conversational AI for foreign language conversation learning, aiming to induce immersion and foster sustained user engagement.<br/>Background: Conversational AI is gaining attention in the field of education for its potential to provide personalized learning and real-time feedback. In foreign language conversation learning through conversational AI, interaction with tutors plays a crucial role in influencing learners&quot immersion and intention to continue. Thus, consideration of anthropomorphic elements that can enhance such interaction is necessary. Anthropomorphism encompasses linguistic elements resembling human language (natural language models, human-like voices, etc.), non-verbal elements (emojis, meme usage, etc.), and visual elements (character images, profiles, etc.), all of which contribute to conveying a sense of social presence. While research integrating anthropomorphic elements into conversational AI is prevalent, there is still a lack of studies specifically focusing on anthropomorphic elements in the context of foreign language conversation learning. Therefore, this study proposes a user interface for conversational AI considering not only human-like appearance but also interaction-related elements such as linguistic and non-verbal factors.<br/>Method: The study focuses on analyzing how the anthropomorphic elements of conversational AI, including visual, linguistic, and non-verbal aspects, influence learners&quot immersion, satisfaction, and intention to continue. Through Focus Group Interview (FGI), the researchers established a visual stage for the AI where learners can naturally engage in conversation, which was then integrated into the prototype. The prototype experiment involved interactions with AIs at varying levels of anthropomorphism: non-anthropomorphic, weak anthropomorphism, and strong anthropomorphism, while participants performed tasks involving art and social-themed conversation scenarios. The key evaluation metrics used in the study were Perceived Ease of Use, Perceived Usefulness, Perceived Interactivity, Perceived Intelligence, Perceived Trust, and Intention to Use, with a focus on participants&quot immersion, learning outcomes, and user satisfaction. <br/>Results: The experimental results showed that the prototype incorporating interactions with a conversational AI featuring strong anthropomorphism scored the highest across all six metrics: Perceived Ease of Use, Perceived Usefulness, Perceived Interactivity, Perceived Intelligence, Perceived Trust, and Intention to Use. This suggests that anthropomorphizing conversational AI, encompassing visual, linguistic, and non-verbal elements, can have a positive impact on foreign language conversation learning. <br/>Conclusion: This study provides evidence that anthropomorphizing the user interface is a crucial factor in enhancing conversation immersion and fostering continued usage intention in the development of foreign language conversation learning services utilizing conversational AI. These findings suggest that in the future design and development of conversational AI, strategic integration of anthropomorphic elements into the user interface should be pursued.<br/>Application: In designing foreign language conversation learning using conversational AI, the user interface can be constructed by considering anthropomorphic elements such as those proposed in this study, including visual, linguistic, and non-verbal aspects of anthropomorphism.