VietTESOL International Convention 2026

The integration of generative artificial intelligence (AI) into education has created new opportunities for language assessment development. In Vietnam, recent educational reforms and changes to the Vietnamese National High School Graduation Examination (VNHSGE) have increased the demand for high-quality reading assessment materials. This study investigates the use of Claude.ai in AI-assisted reading test bank development for Vietnamese Grade 12 students in the EFL context. A descriptive mixed-methods design was employed. A six-stage framework based on structured prompt engineering was developed to generate VNHSGE-oriented reading tasks, including cloze tests, gap insertion tasks, and reading comprehension exercises. The generated materials were evaluated by 21 English teachers and 83 Grade 12 students through questionnaires and open-ended responses. The findings revealed positive perceptions of the AI-generated materials regarding format alignment, clarity, and usefulness for examination preparation. Teachers highlighted the potential of Claude.ai to reduce workload and support assessment design, while students appreciated the practicality of the exercises. However, several limitations were identified, including inconsistent difficulty levels, overly academic vocabulary, and weaknesses in distractor quality. The study concludes that Claude.ai has considerable potential as a supportive tool for reading test bank development in Vietnamese secondary education. Nevertheless, human supervision and expert validation remain essential to ensure assessment quality and pedagogical appropriateness.