World Education Connect Multidisciplinary e-Publication, Vol. VI, Issue VIII (August 2026), pp.1-9
CHRISTINE ANGELIQUE Y. BORJA
Master of Arts in Education, Major in English
Tarlac State University
This quasi-experimental study evaluated the impact of a voice-activated Gemini Live AI intervention on the Oral Communicative Competence (OCC) of Grade 10 Science, Technology, and Engineering (STE) students at Corazon C. Aquino High School during the Fourth Quarter of School Year 2025–2026. Grounded in Canale and Swain’s (1980, 1983) framework, the study utilized a non-equivalent control group pre-test/post-test design involving 66 purposively selected students divided into a conventional instruction control group (n = 32) and an experimental group (n = 34) engaged in daily 10-minute AI conversational sprints over a 25-day period. Oral performance was evaluated using parallel STEM synthesis tasks double-blind scored by four language experts across four sub-competencies (grammatical, sociolinguistic, discourse, and strategic) and a composite Overall OCC level. Baseline equivalence testing via Welch’s t-test confirmed no significant difference in pre-test Overall OCC scores between groups at the "Developing" proficiency level (Control M = 2.900, SD = 0.176; Experimental M = 2.969, SD = 0.312), failing to reject H01 (t(53) = -1.115, p = .2701). Post-intervention inferential analysis using an independent two-sample t-test revealed that the experimental group achieved significantly higher terminal Overall OCC performance (M = 3.173, SD = 0.244) compared to the control group (M = 2.981, SD = 0.213), rejecting H03 (t(64) = -3.407, p = .0011). Within-group paired t-testing confirmed a significant change from baseline pre-test to terminal post-test Overall OCC scores in the experimental group, rejecting H02 (t(33) = -6.303, p < .0001). Sub-competency analysis demonstrated that while the control group achieved higher mean gains in grammatical competence (+0.234 vs. +0.049), the experimental group achieved superior progress in sociolinguistic (+0.201 vs. -0.079), discourse (+0.310 vs. +0.128), and strategic (+0.210 vs. +0.039) competencies. The study concludes that conversational AI functions effectively as an interactive practice engine for overall oral communicative competence rather than a standalone teacher of structural mechanics, validating a hybrid pedagogical model. Recommendations emphasize implementing detailed analytic rubrics per sub-competency for pre-test and post-test assessments, expanding intervention timelines to a six-month duration, and establishing flexible practice environments, encouraging home-based practice for students with personal data or Wi-Fi while establishing dedicated on-campus quiet speaking zones with internet connectivity to support students facing digital access deficits.
Keywords: Oral Communicative Competence, Gemini Live AI, Grammatical Competence, Sociolinguistic Competence, Discourse Competence, Strategic Competence, Hybrid Pedagogy
PUBLICATION GUIDELINES
Publication Ethics and Publication Malpractice Statement
PROCESSING FEES AND INCLUSIONS
