AI tools boost English learners' engagement, a review of 17 studies finds

The authors began with 846 records from a multi-database search and narrowed them through deduplication, screening and full-text eligibility checks. Because the included studies differed substantially from one another, the pooled figure is an average across them, not a fixed effect for every classroom.

Moderator analyses found that the effect varied significantly by the role AI played and by learners' major. The authors say the findings are relevant to researchers, instructors and administrators working in English as a second or foreign language settings.

The study searched multiple databases and found 846 records. After removing duplicates, screening and checking full texts for eligibility, 17 primary studies remained, which together supplied 36 effect sizes for pooling.

The aggregate result showed a large average impact of AI use on engagement among English learners, with g = 0.955 and a confidence interval of 0.720 to 1.191. The authors caution that substantial heterogeneity between studies means this is an average, not a uniform effect.

Moderator analyses tested which variables change the size of the effect. Two emerged as significant: the roles AI plays in learning and the learners' major. The authors frame the results as evidence on AI's effectiveness for engagement in English as a second and foreign language contexts.