Utilitarian Value or Enjoyment Experience? The Formation Mechanism of IELTS Learners’ Continuous Learning Intention in AI-Driven Intelligent Tutoring Systems: Evidence from PLS-SEM, MGA, and ANN

As artificial intelligence becomes increasingly integrated into education, intelligent tutoring systems (ITSs) are emerging as important tools for personalized language learning. However, high learner dropout rates indicate that sustaining continuous intention remains a key challenge. This study examines the mechanisms and differential effects shaping IELTS learners’ continuous intention to use ITSs in listening and reading contexts, using an AI test-preparation assistant and 494 valid responses from 522 Credamo surveys. PLS-SEM results show that perceived usefulness, perceived ease of use, perceived enjoyment, and self-efficacy significantly improve learning attitude, while learning attitude and self-efficacy further strengthen continuous intention. Multi-group analysis reveals contextual differences: perceived enjoyment has a stronger effect in listening (β = 0.340, p < 0.001) than in reading (β = 0.178, p < 0.01), whereas perceived usefulness has a stronger effect in reading (β = 0.424, p < 0.001) than in listening (β = 0.226, p < 0.001). ANN results confirm that enjoyment is the strongest predictor of learning attitude in listening (normalized importance = 100%), while usefulness is the key predictor in reading (normalized importance = 100%). The study extends the TAM by incorporating self-efficacy and perceived enjoyment and highlights the importance of context-specific ITS design for AI-supported language learning.