<![CDATA[This mixed-methods classroom study examined preliminary changes associated with an AI-Augmented Reader’s Theater (AI-RT) model designed to connect English oral performance with learners’ local cultural identity. A quasi-experimental one-group pretest–posttest design was implemented with 64 Grade 5 students (26 boys and 38 girls; approximately 11 years old) in an Intensive English Program at Wat Sra Kaew School, Thailand. The six-week intervention comprised four two-hour sessions per week (48 instructional hours). In groups of four, students developed a four-scene local-culture script, rehearsed orally with ChatGPT, revised their delivery through three to four practice cycles, performed the script, and reflected on the experience. Data were obtained from an analytic oral-performance rubric, a 12-item confidence and cultural-identity questionnaire, and written reflections. The total oral-performance mean increased from 11.80 to 20.25 out of 24. A paired-samples test indicated a significant mean gain of 8.45, t(63) = 9.85, p < .001, 95% CI [6.74, 10.16], with a large standardized paired effect, Cohen’s dz = 1.23. The largest descriptive gains concerned cultural meaning, expressive delivery, and speaking confidence. Questionnaire construct means also increased for confidence (2.85 to 4.20) and cultural identity (2.90 to 4.30). Reflections suggested that repeated, low-pressure AI rehearsal supported willingness to speak, while local stories about Wat Sra Kaew and the Thao Suranari (Ya Mo) monument gave students culturally meaningful content to represent in English. Because the study used one intact cohort without a comparison group, the findings should be interpreted as promising classroom-based evidence rather than definitive causal proof. The study demonstrates how place-based content, performance pedagogy, and guided generative-AI use can be integrated to support meaningful oral participation in primary EFL education.]]>
