Generative Artificial Intelligence (GenAI) has recently attracted growing attention in English as a Foreign Language (EFL) education, offering opportunities for personalised learning and efficiency gains. Yet, alongside these benefits, learners also face psychological tensions, particularly cognitive dissonance, that remain underexplored in current research. Prior work has often examined GenAI use through acceptance-oriented models, which is informative but leaves less room to capture coexisting tensions and the ways learners manage them in practice. Grounded in Cognitive Dissonance Theory, this study investigates how Chinese university students experience and regulate dissonance when using GenAI for English learning. Employing Q methodology, 25 participants were invited to sort 45 statements representing diverse perspectives on GenAI-assisted EFL learning. The Q-sorts were analysed using principal component analysis with Varimax rotation, supported by PQMethod and KADE software, to identify shared patterns of subjective viewpoints. The analysis revealed three salient dissonance types: efficiency-capacity dissonance (concerns over competence loss despite efficiency gains), instrumental-traditional dissonance (conflict between appreciation of AI tools and commitment to teacher-led methods), and trust-reliance dissonance (dependence on GenAI coupled with skepticism about its reliability and ethical risks). Post-sort interviews further identified six self-regulation strategies, including selective neglect, sequencing, reframing, context-based practice, verification, and conformity-based rationalization. The study presents a data-grounded typology of dissonance patterns and a mapping between dissonance configurations and self-regulatory responses, with implications for classroom guidance and for designing GenAI-supported tools that support learners' evaluation and decision-making. • This study uses Q methodology to identify three cognitive dissonance types among Chinese tertiary EFL learners in GenAI-assisted learning. • Six context-specific self-regulation strategies for addressing the dissonance are uncovered via follow-up interviews. • It explicitly distinguishes AI-induced dissonance from traditional education dissonance, extending Cognitive Dissonance Theory. • Practical implications include hybrid instructional models and transparency-focused EdTech design to mitigate AI-related cognitive tension.
Zheng et al. (Mon,) studied this question.