AI-Scaffolded English Oral Performance and Learner Perceptions: A Comparative Case Study of Thai and Chinese University Students
Year: 2026
Type: article
Abstract: This study explored how AI scaffolding helped support English oral communication among Thai and Chinese EFL university students and investigated how these students perceived the usefulness and ease of use of AI tools. Using a comparative mixed-methods analysis, the study purposively selected ten first-year undergraduate students (five Thai and five Chinese participants). The students completed three phases of scaffolded speaking tasks using ChatGPT (for Thai participants) or Doubao (for Chinese participants), along with an extended Technology Acceptance Model (TAM) questionnaire. Speaking performances were analyzed using the Complexity, Accuracy, and Fluency Framework and an Analytic Scoring Rubric, while questionnaire data were examined through descriptive statistics and thematic analysis. The findings revealed that students from both groups improved their vocabulary and idea organization, particularly during structured question-and-answer interactions. By contrast, grammatical accuracy showed little improvement across the three phases. Overall, both Thai and Chinese students viewed AI tools positively and considered them useful and easy to use, although cultural differences influenced what they valued most in the learning experience. The study suggests that generative AI works best as a supportive speaking partner that complements, rather than replaces, human language teachers.
DOI: 10.70730/ddof5466
Language: English
Subfield: Language and Linguistics
Field: Arts and Humanities
Domain: Social Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W7202032597