This study aims to investigate the psychological impact of AI integration on academic self-efficacy, which is self-confidence, especially in the academic field, on the writing performance of English language education students. Using a qualitative approach, this study explores how students who are active as thesis writers to fulfil the requirements of the final project which utilized AI (ChatGPT) as an aid and their efforts in correcting prompts or commands. The data in this study were collected through in-depth interviews and analyzed to understand students' self-evaluative reactions to AI assistance. The findings of this study indicate that ChatGPT increases students' sense of psychological safety and provides efficient support for brainstorming and automatic feedback that can correct thesis drafts. Data processing features in ChatGPT can simultaneously foster a new form of academic dependency because most participants expressed anxiety, doubt, and uncertainty when having to write thesis independently; this indicates a shift in their academic control or locus of control. Furthermore, this suffering also identifies a moral dilemma related to academic integrity, where students actively negotiate the boundaries between utilizing ChatGPT as a legitimate work partner or aid. Besides, students felt guilty due to the use of ChatGPT since it was too instant. This study concluded that ChatGPT is a helpful tool, uncontrolled use risks reducing students' self-confidence, particularly in writing. The study recommended future research using mixed methods and objective performance assessments to evaluate the long-term impact of AI on students' writing competencies
Copyrights © 2026