Quasi-experimental study examines emotional AI's impact on vocabulary learning, revealing nuanced effects based on learner profiles.
Key Points
This study investigates the varied effects of emotional AI on L2 vocabulary acquisition among pre-service teachers, challenging the assumption of universal benefits.
Quasi-experimental study with 147 pre-service teachers interacting with emotional or regular AI agents.
Data collection included pre/post-tests, questionnaires, and AI chat histories.
Quantitative and qualitative analysis to explore interaction patterns and learner archetypes.
No overall difference in vocabulary acquisition between emotional and regular AI agents.
Lower-proficiency learners showed better learning attitudes (p = 0.001) and motivation (p = 0.038) with the regular AI agent.
Exploratory analysis identified four learner archetypes, influencing how emotional AI impacts learning outcomes.