University Students’ Perception of AI in Thesis Writing: Benefits, Challenges and Coping Strategies

Abstract

Writing skills are essential for students in academic, professional, and everyday communication contexts. However, EFL students often face various obstacles in the writing process, such as limited vocabulary, grammatical errors, and difficulty developing ideas systematically. With the development of technology, artificial intelligence (AI) has begun to be widely used to support the writing process through various applications such as ChatGPT, Gemini, Quillbot, Grammarly, and ProWritingAid. This technology can help organize ideas, improve grammar, enhance writing style, and provide immediate feedback, potentially reducing the burden and time of revision. Previous studies from Alharbi (2023), Abimanto (2023), Sumakul et al., (2022), and Schei et al. (2024) have shown that AI is effective in supporting the development of writing skills, providing automatic feedback, and increasing efficiency in the editing process. However, although many studies discuss the effectiveness of AI, there is still a lack of research that deeply explores the perceptions of students, especially undergraduate students, regarding the use of AI in academic writing. This study is based on Robbins & Judge's perception theory and Davis' Technology Acceptance Model (TAM) as a conceptual framework. Therefore, this study was conducted to explore students' perceptions of the use of AI in thesis writing, including the benefits, challenges, and coping strategies they apply. This study uses a qualitative descriptive design with four participants who have used AI during the process of writing their theses. Data were collected through semi-structured interviews with the theory of perceptions and TAM to obtain an indepth picture of students' experiences and views in using AI as a writing tool. These findings are divided into eight themes, the first of which is AI as a Brainstorming Tool, Improve the quality of writing, Negative Impact of Using AI, Limitations of AI in Thesis Writing, AI Results Refinement, Managing the Integration of AI and Self, Role and Guidance of Lecturers, and Institutional Policy. The results show that students view AI as a very helpful tool for developing ideas, understanding difficult concepts, improving grammar, and organizing the structure of their writing. AI is considered capable of increasing efficiency, providing quick feedback, and supporting the thinking process. However, this study also found several challenges, such as dependence on AI, loss of writing originality, lack of confidence, and the risk of receiving inaccurate, fake, or overly formal content. These technical challenges show that AI cannot replace students' critical reasoning and analytical skills. To address these challenges, students employ several coping strategies, such as verifying AI results with academic sources, comparing information through scientific journals, re-editing in their own style, and controlling prompts to make the results more relevant. Students also recognize the importance of maintaining academic integrity by continuing to prioritize independent thinking and avoiding plagiarism. This study provides a comprehensive overview of students' perceptions of the use of AI in thesis writing, as well as revealing the benefits, challenges, and coping strategies that have not been widely discussed in previous studies. These findings provide new insights into how AI plays a role in the academic writing process in higher education. Further research could conduct research that focuses on other aspects of Robbins and Judge's theory, such as attitude, interest, expectations, time, novelty, etc that were not examined in this study. Research using mixed methods can also provide a more comprehensive picture of usage patterns, acceptance levels, and the influence of AI on the process of learning to write in higher education