Improving Young Learners with Copilot: The Influence of Large Language Models (LLMs) on Cognitive Load and Self-Efficacy in K-12 Programming Education Wan Chong Choi Department of Computer Science, Illinois Institute of Technology, US Faculty of Applied Sciences, Macao Polytechnic University, Macao SAR, China wchoi8@hawk.iit.edu Jun Peng* College of Teacher Education, Ningbo University, Ningbo, China School of Education, City University of Macau, Macao SAR, China junpeng@cityu.edu.mo *Corresponding author Iek Chong Choi School of Education, City University of Macau, Macao SAR, China M23091200226@cityu.edu.mo Huey Lei Faculty of Education, The University of Hong Kong, Hong Kong SAR, China hlei@hku.hk Lai Chu Lam School of Informatics, Computing and Cyber Systems, Northern Arizona University, US School of Education, City University of Macau, Macao SAR, China LL2384@nau.edu Chi In Chang Department of Psychology, Golden Gate University, US cchang@my.ggu.edu Abstract—The integration of Large Language Models (LLMs) such as Microsoft Copilot in K-12 programming education has demonstrated the potential to alleviate cognitive load and enhance self-efficacy among young learners. This study examined the impact of Copilot-assisted instruction on cognitive load and self- efficacy in primary school students engaged in programming education. Guided by Cognitive Load Theory and self-efficacy principles, the study employed a quasi-experimental design involving primary school students in Macao. Participants completed pre- and post-tests measuring cognitive load and self- efficacy using the Chinese version of the Cognitive Load Scale (CCLS) and the General Self-Efficacy Scale (GSES). The results indicated significantly reduced students’ cognitive load across the mental load and mental effort dimensions. Concurrently, self- efficacy scores exhibited a statistically significant increase. Correlation analysis revealed a strong negative relationship between cognitive load and self-efficacy, suggesting that students’ confidence in programming tasks improved as cognitive load decreased. These findings highlighted the potential of AI-driven educational tools in optimizing learning environments, reducing cognitive demands, and fostering positive academic self- perception in early programming education. Keywords—Large Language Models, LLMs, Microsoft Copilot, Cognitive Load, Self-Efficacy, Programming Education, AI-Assisted Learning, K12 Learning, Artificial Intelligence in Education I. INTRODUCTION Teachers are adopting diverse approaches and tools in various programming courses, which have been shown to improve learning outcomes [1] [2]. With the rise of Artificial Intelligence (AI), educators increasingly explore AI-driven technologies to enhance teaching and learning processes [3] [4]. Among these innovations, Large Language Models (LLMs) have emerged as potential tools that provide interactive support and real-time feedback, offering new possibilities for student learning [5] [6]. Cognitive load theory (CLT) suggests that learning is most effective when cognitive demands are carefully managed. Zavgorodniaia et al. [7] emphasized measuring cognitive load in educational settings to optimize learning experiences. Programming education often imposes high cognitive demands due to syntax, logic, and debugging complexity. CLT provided a valuable framework for understanding these challenges and was used to theorize the impact of AI-based program generators on learning to program [8]. LLMs like Microsoft Copilot [9] have the potential to reduce cognitive load by offering real-time explanations, adaptive tutoring, and scaffolding strategies. Self-efficacy, an individual's belief in ability to succeed, is crucial to students' motivation and performance. Structured learning environments with cognitive scaffolding have been shown to enhance academic self-efficacy, particularly in mathematics, through targeted instructional strategies [10]. However, research on LLMs’ impact on cognitive load and students' perceived ability to succeed in programming courses remains limited [11]. This study investigated the effects of integrating Microsoft Copilot into primary school programming education, specifically focusing on its impact on students' cognitive load and self-efficacy. We aimed to answer the following research questions. (1) How does using Copilot in K-12 programming education affect students' cognitive load? (2) How does Copilot-assisted instruction affect students' self-efficacy in programming tasks? 284 2025 International Conference on Artificial Intelligence and Education 979-8-3315-2295-7/25/$31.00 ©2025 IEEE 2025 5th International Conference on Artificial Intelligence and Education (ICAIE) | 979-8-3315-2295-7/25/$31.00 ©2025 IEEE | DOI: 10.1109/ICAIE64856.2025.11158328 Authorized licensed use limited to: Macao Polytechnic University. Downloaded on September 25,2025 at 11:05:52 UTC from IEEE Xplore. Restrictions apply.