Original title: Effects of speech-enabled LLM-based chatbots on L2 learning outcomes and affective domains: a multilevel meta-analysis
Eunjin Lee, Hansol Lee, Jang Ho Lee · October 2026
Researchers combined data from 19 studies involving 2,643 language learners who practiced speaking with AI chatbots. These chatbots listen to spoken words and talk back using modern language models. The analysis tested whether voice practice improves learning and emotional factors like confidence. They found strong, positive effects on both learning outcomes and student attitudes. However, the exact impact depends on things like the student's skill level, their school grade, and the specific setup of the practice.
Learning to play tennis against a wall can feel repetitive, and finding a partner at your exact skill level is tough. When someone practices against a wall, they hit the ball, but they never get realistic, varied returns that feel like a real match. In this paper, speech-enabled chatbots act like an adaptable training partner that actually hits the ball back in real time. The researchers gathered results across many different practice courts to see if hitting back and forth with these virtual partners genuinely builds real-match skill and player confidence.
Practicing conversation in a new language can be intimidating and hard to schedule with real people. This synthesis shows that talking out loud with AI chatbots actually boosts both learning results and student confidence, though the benefits vary depending on who uses them and how.
The findings are constrained to the 19 studies and 2,643 participants analyzed, and the abstract does not report long-term retention or identify which specific types of chatbots or LLM architectures worked best.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Effects of speech-enabled LLM-based chatbots on L2 learning outcomes and affective domains: a multilevel meta-analysis — Eunjin Lee, Hansol Lee, Jang Ho Lee — IRAL - International Review of Applied Linguistics in Language Teaching — doi:10.1515/iral-2026-0172
Abstract licence: CC BY
Explanation generated by Synthesis from the abstract — it is not the original paper.