Abstract
The integration of AI in language learning is gaining popularity through Intelligent Computer-Assisted Language Learning (ICALL), facilitating personalized learning environments. However, existing ICALL solutions for Second Language Acquisition (SLA) suffer from limited effectiveness due to outdated AI practices. To address this, we propose the utilization of generative AI to overcome such limitations and create personalized learning experiences. This paper presents our initial prototype that employs various generative AI models to construct a virtual scene with a conversational agent for linguistic practice, guided by a set of design principles. Through a user-centered design process with SLA learners, we iteratively refine the prototype and produce design guidelines. We discuss the challenges of using the prototype as a standalone application for linguistic training and provide recommendations for integrating generative AI in SLA. Our paper aims to establish a foundation for generative AI approaches in SLA and offer design considerations for AI-based language learning tools.
