Not a universal benefit: Examining the differential effects of emotional AI on L2 pre-service teachers' language learning

Synthesis: This study challenges the assumption that emotional design in educational AI provides universal benefits, investigating when, for whom and how it impacts L2 vocabulary learning. A quasi-experiment with 147 pre-service teachers found no overall difference in vocabulary acquisition, but a nuanced pattern emerged: the regular agent preserved significantly better learning attitudes and intrinsic motivation for learners with lower baseline proficiency. Qualitative analysis identified four distinct learner archetypes, suggesting emotional AI efficacy may be moderated by these profiles.

Key Findings

This study challenges the assumption that emotional design in educational AI provides universal benefits, investigating when, for whom and how it impacts L2 vocabulary learning. A quasi-experiment with 147 pre-service teachers found no overall difference in vocabulary acquisition, but a nuanced pattern emerged: the regular agent preserved significantly better learning attitudes and intrinsic motivation for learners with lower baseline proficiency. Qualitative analysis identified four distinct learner archetypes, suggesting emotional AI efficacy may be moderated by these profiles.

Connected Concepts

Connected Articles

Citation## Citation

Wang, Z., & Pang, H. (2026). Not a universal benefit: Examining the differential effects of emotional AI on L2 pre-service teachers' language learning. British Journal of Educational Technology.