Enhancing Metaliteracy Competence through a ‘Learner-Centered Topic-Determination Model’ in Generative AI-Assisted Discussion Classes : A Case Study of a READING CLASSICS OF GREAT LITERATURE Course Using Brave New World

Generative artificial intelligence is now embedded in everyday academic and professional practices. Higher education accordingly faces a structural challenge: how to provide learners with procedural experiences through which they analyze the propositional structure of AI-generated outputs, evaluate the credibility of evidence and the validity of reasoning, and reconstruct these elements as their own arguments. This study examines the implementation of the required general education course READING CLASSICS OF GREAT LITERATURE at Pusan National University in Busan. The course employs what this study designates as A Two-Stage Reflective Problem-Framing Process, consisting of learner-directed initial topic selection, a first round of discussion, critical review of identified problems, reconstruction of the topic, and a second round of discussion. This study argues that the process functions as an educational mechanism for fostering learners’ metaliteracy competencies. In the first stage, the use of generative AI was deliberately restricted so that learners could formulate their own critical awareness of the text. In the second stage, the university’s integrated generative AI service, Sanjini AI, together with group-specific customized GPTs, was employed as a mediating tool for learners’ self-assessment and reflective revision. This study conducted a qualitative content analysis of reflective reports on ‘second-stage topic revision and conceptual renegotiation’ written by 52 students enrolled in Sections (가), (나), and (다) during the first semester of the 2026 academic year. The analysis reveals that learners articulated the limitations of their initial discussion topics in their own terms, redefined key concepts in operational ways, recalibrated the relationship between external evidence and internal textual interpretation, and employed AI-generated outputs as mediating resources for verifying their own reasoning. The study concludes that the core educational outcome of a learner-centered topic-selection curriculum is enabling students to recognize, through sustained discussion, the significance of logical reasoning competencies and their relevance to the age of generative AI.