Yi-Ru Her
Department of Computer Science, University of Taipei
E-mail: m11316003@go.utaipei.edu.tw
Ah-Fur Lai
Department of Computer Science, University of Taipei
E-mail: laiahfur@gmail.com
Yen-Hung Chen
Department of Computer Science, University of Taipei
E-mail: yhchen@utaipei.edu.tw
Abstract
This action research study explores the application and effectiveness of generative AI tools in elementary school programming and computational thinking education. The research involved 108 sixth-grade students from four classes in an elementary school in Taipei City, randomly divided into experimental and control groups. The experimental group used ChatGPT combined with constructed response items for micro:bit programming learning, while the control group followed traditional demonstration-practice teaching methods. Students were initially classified into high, middle, and low-ability levels based on their performance in the Taiwan International Computational Thinking Challenge pretest to observe the impact of AI integration on students of different abilities.
The results revealed that AI tools had differential effects on students across ability levels: high-ability students effectively utilized AI to explore innovative functions and expand their learning boundaries; middle-ability students received timely support through AI, improving learning efficiency; low-ability students struggled to use AI independently and remained highly dependent on teacher guidance. Programming projects from the experimental group demonstrated greater innovation and better structural organization than those from the control group, though this advantage was primarily reflected among high-ability and some middle-ability students.
This study proposes a staged approach to AI implementation in teaching, emphasizing the importance of the teacher’s role transformation from knowledge transmitter to learning designer and facilitator, while exploring new possibilities for adaptive learning through AI integration. The findings provide empirical references for differentiated instruction in elementary programming education and indicate directions for future research and practice in AI-integrated education.
Keywords: Generative Artificial Intelligence, Computational Thinking, Action Research, Adaptive Learning, Constructed Response Items