Bridging AI and Human Expertise in ESP: Optimizing AI-Integrated Instruction for Job Application Skills

This study investigates the integration of Artificial Intelligence (AI) tools within a 15-week English for Specific Purposes (ESP) course designed to enhance job application competencies among senior-level university students (n=11) in Korea. Adopting a mixed-methods action research approach, the study explores how AI-generated feedback from tools such as ChatGPT, Clova Note, and YouGlish, combined with explicit teacher scaffolding, impacts learners’ résumé writing, cover letter drafting, and interview skills. Quantitative analyses revealed significant improvements in students’ lexical accuracy, structural coherence, and oral proficiency, albeit with notable variability reflecting individual proficiency differences. Qualitative insights indicated that while learners appreciated AI’s efficiency and iterative feedback capabilities, lower-proficiency students struggled with the interpretation and application of AI suggestions, underscoring the necessity for targeted human mediation. Ethical concerns surrounding data privacy and AI’s limited adaptability to specialized professional contexts also emerged, highlighting critical areas for further pedagogical attention. Based on these findings, the study advocates a balanced AI-human instructional model, providing concrete pedagogical strategies for proficiency-sensitive AI integration, ethical AI use, and structured peer collaboration. These recommendations offer valuable implications for educators aiming to harness AI’s potential effectively while maintaining essential human oversight in ESP instruction.