Built around how learning works.

Finding

A meta-analysis of 48 second-language-learning experiments found medium-to-large benefits from spaced practice.

What it means for Lessonling

Practice is spread across the week rather than packed into the lesson: a short exercise on the student's practice days, at a time that suits them, instead of one long homework session. We do not claim that any one schedule, daily or otherwise, is the best.

Source

Kim, S. K., & Webb, S. (2022). The Effects of Spaced Practice on Second Language Learning: A Meta-Analysis. Language Learning, 72(1), 269–319. https://doi.org/10.1111/lang.12479

This study did not test Lessonling.

Finding

Learning is strengthened when you actively recall and use information instead of only looking at it again.

What it means for Lessonling

Exercises ask the student to produce: recall a word, build a sentence, answer in a conversation. The problem areas and what Charlie has learned about the student come back in later exercises when Charlie picks the next one, and words from the student's list are drawn into them, so they are retrieved again rather than re-read.

Source

Carpenter, S. K., Pan, S. C., & Butler, A. C. (2022). The science of effective learning with spacing and retrieval practice. Nature Reviews Psychology, 1, 496–511. https://doi.org/10.1038/s44159-022-00089-1

This review did not test Lessonling.

Finding

A classroom meta-analysis found significant and durable effects from oral corrective feedback.

Study detail

Effects were larger for prompts than for simply recasting the correct answer.

What it means for Lessonling

Feedback arrives while the learning is happening, not a week later. In short exercises Charlie can first prompt the student to fix the mistake themselves; the full correct sentence and one explanation follow. In conversations Charlie avoids interrupting and nudges at most once per turn, and keeps corrections for the debrief.

Source

Lyster, R., & Saito, K. (2010). Oral Feedback in Classroom SLA: A Meta-Analysis. Studies in Second Language Acquisition, 32(2), 265–302. https://doi.org/10.1017/S0272263109990520

This meta-analysis studied classroom feedback; it did not test Lessonling.

Finding

A meta-analysis of 28 studies found a positive overall effect of chatbot-assisted language learning.

Study detail

Effectiveness varied with factors including educational level, language level, interface design and interaction capability. AI conversation can support language learning – but how it is designed matters.

What it means for Lessonling

We treat design as the point: Charlie works inside Telegram, at the student's level, from the teacher's plan, with fixed rules for difficulty and correction.

Source

Wang, F., Cheung, A. C. K., Neitzel, A. J., & Chai, C. S. (2025). Does Chatting with Chatbots Improve Language Learning Performance? A Meta-Analysis of Chatbot-Assisted Language Learning. Review of Educational Research, 95(4), 623–660 (published online 2024). https://doi.org/10.3102/00346543241255621

This meta-analysis did not test Lessonling.

Finding

A meta-analysis of 41 studies found a moderate-to-large positive effect of generative-AI chatbots on second-language acquisition.

What it means for Lessonling

Generative AI can hold a real conversation at the learner's level, so Charlie offers conversation and role-play, not only one-line drills.

Source

Li, M., Wang, Y., & Yang, X. (2025). Can Generative AI Chatbots Promote Second Language Acquisition? A Meta-Analysis. Journal of Computer Assisted Learning, 41(4), e70060. https://doi.org/10.1111/jcal.70060

This meta-analysis did not test Lessonling.

Finding

Research on AI-supported speaking practice has reported increased willingness to communicate and reduced foreign-language anxiety in some learner groups.

What it means for Lessonling

This study is one reason Charlie invites voice messages and short spoken answers: a private place to speak, make mistakes and try again. It followed one group of learners over six weeks, so we read it as encouraging rather than conclusive.

Source

Artificial intelligence in EFL speaking: Impact on enjoyment, anxiety, and willingness to communicate. (2024). System, 121, 103259. https://doi.org/10.1016/j.system.2024.103259

This study involved one group of university learners; it did not test Lessonling.

Finding

A 2026 study involving 89 English learners found significantly higher behavioural, cognitive and social engagement when GenAI speaking practice included structured teacher guidance.

What it means for Lessonling

This is the closest match to how Lessonling works: the teacher sets the direction, the AI runs the practice. Charlie works from the teacher's plan, and the teacher sees what happened before the next lesson.

Source

Min, S., Gao, Y., & Zhang, J. (2026). Exploring L2 learner engagement in spoken English practices with a GenAI-powered chatbot. System, 138, 103982. https://doi.org/10.1016/j.system.2026.103982

This study measured engagement with a different chatbot; it did not test Lessonling.

Guidance

UNESCO advocates a human-centred approach to generative AI in education, with meaningful pedagogical design, human agency and appropriate human oversight.

What it means for Lessonling

AI should extend a teacher's work, not replace the teacher. Charlie never decides what a student learns, and its observations stay separate from what the teacher entered.

Source

UNESCO. Guidance for Generative AI in Education and Research. https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research

This is general guidance, not a study of Lessonling.