Examining Learning Experiences in a Process-Based Liberal Arts English Course Integrating ChatGPT : A Case Study of a Writing-for-Speaking Transfer-Based PAIL Model

This study examines learning experiences in a process-based liberal arts English course integrating generative AI (ChatGPT) through the application of a Writing-for-Speaking transfer-based Process-based AI-mediated Language Learning (PAIL) Model. The PAIL Model is an instructional framework that structures drafting, AI-mediated feedback, revision, reflection, and speaking performance as a continuous learning process and was implemented in a university-level English speaking course. The study was conducted during the second semester of the 2025 academic year with 22 undergraduate students. Data were collected through an end-of-semester survey consisting of 15 items and reflective learning journals produced during the learning process. The results indicate that learners experienced improvements in cognitive performance, including enhanced comprehension, refined expression, and increased error awareness, as well as positive affective changes such as increased motivation, participation, and performance confidence. These findings suggest that AI support functioned as interactive scaffolding within learners’ zones of proximal development (ZPD) in process-based tasks. By conceptualizing generative AI as a mediator of learning processes rather than an outcome-generating tool, this study offers pedagogical implications for process-centered instructional design in AI-integrated liberal arts English education.