K-NN Based Prediction of AI Tool Utilization by Non-Technical University Students

Abstract

The increasing integration of Artificial Intelligence (AI) tools, particularly ChatGPT, into higher education necessitates a deeper understanding of their adoption patterns among non-technical students. While AI offers significant benefits for learning and academic tasks, its utilization varies across disciplines, with non-technical fields often exhibiting lower adoption rates. This study addresses the critical need to predict AI adoption among students in non-technical majors such as Business, Education, Humanities, and Social Sciences. We employ the K-Nearest Neighbor (K-NN) algorithm to classify and forecast the likelihood of these students using ChatGPT for academic purposes. The dataset, comprising survey responses from 48 non-technical students, includes attributes like AI knowledge level, frequency of personal and academic AI use, and interest in AI careers. After rigorous data preprocessing, including encoding and normalization, the dataset was split into training (70%) and testing (30%) sets. The K-NN model, with an optimized K-value determined through cross-validation, utilized Euclidean distance for classification. Our findings indicate that approximately 39.6% of non-technical students are predicted to utilize AI tools like ChatGPT for their academic activities, closely aligning with actual survey responses. This research provides valuable insights for educational institutions to tailor teaching methods, offer targeted support, and develop relevant digital literacy programs, ensuring AI becomes an inclusive and empowering educational tool for all students.