The effects of generative AI on foreign language emotions: A meta-analysis

Area of Science:

• Educational Technology
• Applied Linguistics
• Artificial Intelligence in Education

Background:

• Growing interest in Generative Artificial Intelligence (GenAI) in second language education.
• Lack of systematic evidence on GenAI's impact on foreign language (FL) emotions.

Purpose of the Study:

• To systematically synthesize evidence on the effects of GenAI-assisted instruction on FL emotions.
• To explore how GenAI influences both positive and negative emotional states in language learning.

Main Methods:

• Meta-analysis of 23 experimental and quasi-experimental studies.
• Inclusion of 43 effect sizes from 3461 participants.
• Application of a three-level random-effects model.

Main Results:

• GenAI interventions show a large, significant positive effect (g = 0.788) on FL emotions.
• Improvements were more pronounced in reducing negative emotions (anxiety, boredom) than enhancing positive ones (enjoyment).
• Limited evidence found for moderators related to technological, pedagogical, and social affordances.

Conclusions:

• GenAI-assisted language learning holds significant emotional potential.
• GenAI is particularly effective in alleviating negative emotional states.
• Emotionally supportive AI-mediated FL instruction is valuable for learners.