Answers are grounded in this paper and cite their sources.
Systematic review evaluates chatbots' impact on oral proficiency in language learners, suggesting effective applications and challenges.
Developing oral proficiency in English as a foreign language remains a persistent challenge, largely due to limited opportunities for authentic and low-stress speaking practice. Chatbots provide learners with accessible conversational partners, yet their role in improving specific dimensions of spoken performance has not been comprehensively synthesised. While earlier reviews focused on motivation or technology adoption, no study has systematically examined how chatbot-mediated interaction affects the core qualitative features of oral proficiency. This systematic review evaluates the effects of human–chatbot interaction on six features of oral production: range, accuracy, fluency, interaction, coherence, and intelligibility. It also examines learner perceptions and highlights technological and pedagogical challenges. Following PRISMA guidelines, a search of Scopus and Web of Science identified ten empirical studies published between 2021 and 2024 that assessed the impact of speech-based or multimodal chatbots on English oral proficiency. Methodological quality was assessed using the Mixed Methods Appraisal Tool. Findings revealed the strongest gains in intelligibility and lexical range, with moderate improvements in fluency and accuracy, particularly among beginner and intermediate learners. Interaction and coherence were rarely evaluated, underscoring significant research gaps. Learners generally reported higher motivation, reduced anxiety, and greater willingness to communicate, although limitations such as narrow topic scope, speech recognition errors, and non-human-like responses restricted authenticity. Overall, chatbots show strong potential for oral language development, though their effectiveness depends on design quality, implementation context, and alignment with communicative teaching principles.
No takes yet. Share an insight, caveat, or question.
Soriano-Gonzalez et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: