Academic writing poses significant challenges for students, particularly non-native English speakers. While existing tools, such as Grammarly, provide surface-level corrections, they often lack detailed explanations, long-term skill development and personalized support. In this paper, we introduce WILLM, a system for Academic Writing Improvement based on Large Language Models (LLMs). WILLM provides context-aware feedback on grammar, vocabulary, coherence, and organization while integrating active recall quizzes and personalized reviews to enhance long-term writing proficiency. A three-week user study with 19 non-native English-speaking participants demonstrated improvements in grammar and vocabulary scores and high usability ratings (SUS score: 84.21). Results suggest that further refinement of long-term improvement features is essential for enhancing academic writing development.
Citation:
Mou, Y., Fathi, F., Thillen, B., & Decker, S. (2025). WILLM: A system for academic writing improvement based on large language models. In A. I. Cristea, E. Walker, Y. Lu, O. C. Santos, & S. Isotani (Eds.), Artificial intelligence in education (AIED 2025, Lecture Notes in Computer Science, Vol. 15881). Springer. https://doi.org/10.1007/978-3-031-98462-4_5