Clue before correction: ChatGPT-enhanced strategy for promoting autonomous and reflective language learning

Purpose This study presents a generative AI-supported revision task designed to promote adaptive learning, learner autonomy, and metacognitive engagement in second language (L2) writing through clue-based interaction with ChatGPT. Unlike traditional chatbot use for direct error correction, which positions learners as passive recipients, this activity was intentionally structured to require active problem-solving, prompting students to infer correct solutions from guided hints.Design/methodology/approach Fifty-eight university students (CEFR A1–B1) completed a bilingual activity in English or Spanish that used indirect prompts to guide self-revision, supported by ChatGPT. Learner–AI interaction transcripts and open-ended reflections were analyzed using Cognitive Load Theory, formative feedback models, and the newly developed ICAP-ME framework, which extends the ICAP model to include metacognitive and affective dimensions.Findings ChatGPT provided structured, adaptive feedback that reduced cognitive load and supported personalized revision pathways, particularly for more proficient learners. Learner prompting behavior strongly influenced feedback quality and depth.Originality/value This study introduces (1) a replicable instructional design for GenAI-supported revision, (2) the ICAP-ME pedagogical framework for analyzing AI-mediated learning, and (3) design principles for equitable, developmentally appropriate use of GenAI in language education. It contributes to emerging research on how AI tools can foster autonomy, reflective engagement, and strategic learning.