A Study on Business Students’ Use of Large Language Models to Learn Programming Languages: The Case of University A

Abstract

In recent years, the use of Large Language Models (LLMs) to assist learning has become increasingly widespread among university students, particularly in interdisciplinary learning scenarios. This study examines the behavioral patterns of business students at University A using questionnaire surveys, in-depth interviews, and K-Means clustering analysis to explore their use of LLMs for learning programming languages. It ultimately identifies and refines three distinct behavioral profiles: Deeply Engaged Learners, Tool-Assisted Learners, and Passive-Dependent Learners. Building on this foundation, the study systematically analyzes the learning motivations, behavioral characteristics, and core pain points of each group. It proposes optimization strategies for Human-LLM Collaboration across four dimensions—scenario adaptation, role division, interaction design, and knowledge internalization—to provide practical guidance for business students’ programming language learning pathways in a technology-empowered context.