Tesis
Efl postgraduate learners\' perception towards quillbot utilization in academic writing / Mutiara Khofifah
Abstrak
This mixed-methods study investigated how EFL (English as a Foreign Language) master s degree students at Universitas Negeri Malang perceived and utilize QuillBot an AI-powered writing assistance tool in their academic writing. Using the Technology Acceptance Model (TAM) as its theoretical framework the research examined students perceptions across three dimensions perceived usefulness perceived ease of use and behavioral intention to use while also exploring their experiences and reasoning behind these perceptions. The study employed an explanatory sequential design involving 72 master s degree students in English Language Education. The participant were selected using proposive sampling the participant were chosen based on specific criteria. Data was collected through a 17-item close-ended questionnaire based on TAM focus followed by semi-structured interviews to provide deeper insights on students rsquo perception and the reason behind them. The research aimed to understand not only what students thought about QuillBot but also why they held these perceptions and how these influenced their intentions to continue using the tool. The quantitative results revealed generally positive perceptions across all three TAM dimensions though with significant variations among students. Perceived usefulness achieved the highest mean score (3.157) with students particularly valuing QuillBot s time-saving benefits (M 3.222) its utility for academic writing needs (M 3.263) and quality improvement capabilities (M 3.138). However 40.3% of students expressed disagreement on multiple usefulness items indicating mixed feelings about the tool s actual benefits. Perceived ease of use scored positively (M 3.086) with students finding QuillBot technically simple to operate (M 3.388) and its interface understandable (M 3.097) though flexibility and adaptability received lower ratings (M 3.000). The most concerning finding was in behavioral intention to use which achieved the lowest score (M 2.724) among the three dimensions with 55.6% of students recording multiple low scores on intention items particularly regarding regular usage (M 2.597) and increasing usage over time (M 2.430). The qualitative findings provided crucial explanations for these quantitative patterns revealing sophisticated student reasoning that extended far beyond simple utility calculations. Students with positive perceptions described strategic selective usage patterns while maintaining critical evaluation of QuillBot outputs. They valued the tool for specific functions like paraphrasing vocabulary enhancement and making their writing more natural and academic but emphasized the importance of not relying on it completely. Conversely students with negative perceptions articulated comprehensive concerns about output quality contextual appropriateness and academic integrity. They frequently reported that QuillBot s suggestions were awkward contextually inappropriate or required extensive revision that negated any efficiency benefits leading them to completely avoid the tool for serious academic work such as theses and journal articles. A central theme that emerged from the interviews was the dependency dilemma where students across all perception categories expressed concerns about becoming overly reliant on QuillBot and potentially undermining their independent writing skills and critical thinking development. Even students who acknowledged current benefits often described evolving awareness of potential risks and planned to reduce their usage over time. This tension between short-term utility and long-term learning concerns significantly influenced behavioral intentions explaining why students could recognize QuillBot s usefulness while simultaneously expressing reluctance to continue using it. Additional factors influencing acceptance included academic integrity concerns with students worried about maintaining the authenticity of their academic voice and context sensitivity where the tool was deemed more appropriate for preliminary work but inadequate for sophisticated academic arguments requiring disciplinary expertise. The study s major conclusion reveals that academic technology acceptance involves complex evaluations extending beyond traditional TAM factors to include learning objectives professional development needs and ethical considerations. The research identified that master s students as emerging academic professionals held higher standards and greater concerns about maintaining scholarly integrity compared to undergraduate students. The findings suggest that while AI writing tools like QuillBot may offer certain benefits for academic writing support their integration into educational contexts requires careful consideration of pedagogical objectives and long- term learning impacts. The study recommends that educators emphasize strategic supplementary usage rather than primary reliance on such tools while institutions should develop comprehensive AI literacy programs and evidence-based policies that balance technological assistance with authentic learning experiences.