Programming Learning Supported by Generative Artificial Intelligence: Student Programming Performance and Cognitive Load

• In Event: Virtual Poster Session
  In Virtual iPresentation Only: Division C - Section 1e: Engineering and Computer Science Virtual Poster Session

Abstract

The complexity of learning programming requires students to master syntax, logic, and debugging tasks, which can place a high cognitive load on them and lead to academic failure. To address this gap, this study recruited 76 university students to participate in the controlled experiment of programming learning. There are two experimental groups (EG1: Supported by generative artificial intelligence; EG2: Supported by a human teacher), and one control group (CG: No support). The results showed that the group that received support from generative artificial intelligence in programming learning reported more favorable results in improving programming performance and decreasing levels of cognitive load. It provides practical and theoretical insights into the educational practice of GenAI-supported programming.

Authors

• Heng Zhang, University of Hong Kong

• Jing Zhang, Universiti Sains Malaysia