AI ASSISTANCE OR OVERRELIANCE? ANALYZING ESP STUDENTS’ CRITICAL USE OF GENERATIVE AI

Abstract

The rapid integration of generative artificial intelligence (GenAI) tools as reshaped how English for Specific Purposes (ESP) students —like in engineering disciplines— approach language production. Despite their growing presence, limited research has examined the risk of overreliance on GenAI and its potential to diminish students’ critical language awareness. Using a purposive sampling in which 27 students who met the criteria, the study addressed three questions: (1) To what extent do ESP students critically evaluate AI-generated suggestions? (2) What factors influence their acceptance, revision, or rejection of AI output? and (3) In what ways do ESP students utilize AI tools to support their English skills? A qualitative approach was utilized with the data collected through class observation and open-ended questionnaires. The findings revealed that while AI use is widespread, many students demonstrated critical engagement, often modified sentence structures to sound more humanized. However, lack of confidence in their English abilities and time constraints were key reasons for passive acceptance of AI output. Students used AI mostly for overcoming their lexical barriers. These findings underscore the need for careful integration of GenAI into ESP classrooms, supported by targeted instruction and thoughtful lesson design, particularly for students with lower proficiency. Future research should further examine patterns of overreliance across more diverse learner populations.