| Sumario: | The rapid advancement of artificial intelligence has led to the increasing use of AI-powered tutors in English as a Foreign Language (EFL) instruction, particularly for speaking skills development. Despite growing adoption, empirical evidence comparing the effectiveness of AI-powered tutors and human tutors in developing oral fluency remains limited and inconclusive. This study aims to comparatively examine the efficacy of AI-powered tutors and human tutors in enhancing EFL learners' oral fluency across multiple fluency dimensions. This research employed a comparative experimental design involving two groups of EFL learners who received equivalent instructional content over a fixed intervention period, differentiated only by tutoring modality. One group engaged in AI-powered tutoring, while the other received instruction from human tutors. Pretest and posttest measures of oral fluency were administered, focusing on temporal, interactional, and discourserelated indicators. Quantitative data were analyzed using inferential statistical techniques to determine within-group and between-group differences. The findings reveal that both AI-powered tutors and human tutors significantly improved learners' oral fluency. AI-powered tutors were particularly effective in enhancing temporal fluency, including speech rate and reduced hesitation. Human tutors demonstrated superior effectiveness in developing interactional competence, discourse coherence, and pragmatic appropriateness. The study concludes that AI-powered tutors serve as effective supplementary tools for oral fluency practice, while human tutors remain essential for higher-level communicative development. A hybrid instructional approach is recommended to maximize EFL speaking outcomes.
|
|---|