Evaluating the Educational Impact of AI-Generated Test Cases in a Test-Driven Development Environment for Python Beginners

Test automation and Test-Driven Development (TDD) have emerged as key strategies for improving code quality and learning effectiveness, while artificial intelligence tools such as GPT are increasingly applied to automated test case generation. However, simply increasing the number of test cases or relying on automation tools does not guarantee sufficient fault detection performance, and empirical evidence of their educational effectiveness for novices remains limited. This study implements a TDD-based practice environment for Python beginners and compares researcher-designed cases, GPT-generated cases, GPT with strategy-guided prompts (GPT+Strategy), and student-written cases. Results show that GPT+Strategy cases achieved higher fault detection and efficiency than even the researcher-designed cases. TDD-based practice strengthened learners’ logical thinking and error recognition skills. These findings provide practical implications for programming curriculum design, AI-assisted learning tool development, and software quality assurance education.