Design principles for supporting self-regulation and bilingual competence through AI-facilitated narrative exploration in university students

Abstract

University students increasingly need self-regulated learning (SRL) and Chinese-English bilingual competence, yet many technology-enhanced language learning environments remain tool-centred and provide limited integration of culturally situated practice, adaptive strategy support, and metacognitive reflection. This conceptual design-principle study proposes the DeepSeek-Facilitated Dual-Situated Narrative Model (D-DSNM), which brings together SRL theory, the Dual-Situated Learning Model (DSLM), and AI-facilitated narrative exploration. The framework was developed through a critical literature synthesis and theory-driven conceptual analysis rather than through empirical testing. SRL phases and DSLM situations were first mapped deductively; findings from AI-assisted language learning and technology-enhanced SRL were then used as analogous evidence to formulate three proposed mechanisms: cognitive conflict to metacognitive awareness, adaptive scaffolding and strategy activation, and cyclical reinforcement and strategy internalization. These mechanisms were translated into five theoretically derived design directives with sample prompts and classroom scenarios. Within the model, DeepSeek is treated as a proposed technical resource for generating and adapting bilingual narratives and reflective prompts, whereas anticipated learning outcomes are attributed to the instructional design that sequences narrative exploration, reflection, feedback, and learner choice. The D-DSNM is intended to clarify how culturally embedded bilingual narratives could be coordinated with SRL processes in university education. Its contribution is therefore propositional: the mechanisms, directives, and projected outcomes require empirical validation across learners, instructors, AI platforms, and educational contexts.