The Evolution of Learners' Interaction, Cognition, and Emotion in AI Digital Human-Mediated Learning

This study explores how learners' interactional relationships, cognitive engagement, and emotional responses evolve through sustained interaction with an AI digital human in a vocational English writing course. A qualitative case study was conducted over six weeks, during which twelve students engaged regularly with an AI agent named Ava that provided feedback, prompts, and revision support. Data were collected from semi-structured interviews, weekly reflection journals, and system interaction logs. Thematic analysis revealed three core developmental trajectories: (1) Interactional Repositioning, where learners shifted from viewing Ava as a tool to perceiving her as a dialogic partner; (2) Emotional Realignment, marked by a transition from discomfort or skepticism to trust and empathy; and (3) Cognitive Reframing, where learners moved from surface-level editing to metacognitive writing awareness. These findings suggest that AI digital humans can play a relational role in shaping learners’ emotional receptiveness and strategic thinking, particularly when interactions are consistent and socially embedded. The study contributes to a learner-centered understanding of AI-mediated learning and highlights the need to consider affective and cognitive development as intertwined processes in AI-supported education.