Student Academic Writing Analysis for AI-Based Adaptive Learning System Development, Case Study: Building Engineering Education Students

The advancement of Artificial Intelligence (AI) technology in higher education has been suboptimal in tackling the challenges of student academic writing, as current systems primarily concentrate on technical enhancements and have yet to effectively aid students in formulating logical arguments and coherently organizing ideas. This research seeks to evaluate students' academic writing abilities to inform the development of an AI-driven learning system tailored to their specific requirements. A descriptive quantitative methodology was employed to gather data via online questionnaires from 50 students enrolled in the Building Engineering Education Study Program. The instruments were predicated on five dimensions: writing structure, language proficiency, argumentation and logic, reference utilization, and technical writing skills, utilizing a 1-5 Likert scale. The findings indicate that PTB students fall within the "Very Capable" classification, attaining an average score of 82%. The utilization of references garnered the greatest score at 87%, although writing structure and logical argumentation remain at the lowest level, scoring 79%. The findings reveal difficulties in shifting from descriptive technical writing to argumentative academic discourse, underscoring the necessity for an AI-driven learning system that incorporates a corpus-based instructional method to deliver comprehensive feedback on the formation of coherent argumentation structures. The study advocates for the creation of a comprehensive AI system prototype and its extension to other engineering programs to enhance the academic literacy of students across the nation.