Using the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) framework, this descriptive qualitative study explores how final-semester students use ChatGPT in English learning, the ethical challenges they face, and how they balance the tool’s convenience with academic honesty during thesis writing. Purposive sampling was used to select sixteen eighth-semester students from the English Education Study Program at UINFAS Bengkulu. Data were gathered through semi-structured interviews, direct observations, documentations, and then analysed using thematic analysis. The study reveals three main points: first, driven by practical benefits, ease of use, and daily habits, ChatGPT serves as an instant language assistant for brainstorming and checking grammar, despite financial costs for premium features; second, students experience serious worries, including anxiety about plagiarism detection, the risk of becoming mentally lazy, and distrust toward ChatGPT’s incorrect responses, making formal disclaimers about its use secondary to verifying the actual research sources; third, to solve these problems, students take active control by double checking ChatGPT’s answers through Google Scholar and rewriting texts using manual editing and paraphrasing tools to keep their original writing style. This study concludes that eighth-semester students do not blindly accept ChatGPT’s outputs but use it wisely as a starting framework, keeping human intelligence as the main driver in writing.