AI-Assisted Language Assessment: Basis for Senior High School Language Arts Instruction Assessment

This study investigated the role of AI-assisted language assessment in evaluating students' English-speaking proficiency, focusing on grammar, fluency, pronunciation, and vocabulary. It aimed to determine how AI-generated feedback can serve as a basis for enhancing language arts instruction assessment. The study involved seventy (70) Grade 11 HUMSS students from ACLC College of Ormoc City, Inc. A descriptive research design was employed, and students’ speaking performances were assessed using an AI tool. Scores were analyzed based on the Common European Framework of Reference for Languages (CEFR), identifying the dominant proficiency levels for each linguistic component. Results revealed that the majority of students achieved high proficiency levels in grammar and vocabulary, predominantly reaching C1 and C2 classifications. However, greater variability was observed in fluency and pronunciation, where some students exhibited occasional hesitations and minor inconsistencies. Despite providing structured and standardized evaluations, the AI tool demonstrated limitations, such as generating vague, repetitive feedback and displaying difficulty in adapting to varied discussion themes. Issues in the clarity and accuracy of the quantitative scores were also noted. Moreover, the AI tool was observed to consistently assign overly high scores across linguistic competencies, raising concerns about its reliability as an assessment measure. As a result, the study does not recommend the use of AI as a sole assessment method for speaking proficiency at the Senior High School level, emphasizing the continued need for human evaluation to ensure more accurate, fair, and meaningful feedback.