A Study on the Impact of AI-Assisted Project-Based Learning Design on Innovation Ability

Abstract

This study explores the impact of integrating generative artificial intelligence (AI) into Project-Based Learning (PBL) on the innovation ability of undergraduate students in data science education. Utilizing a quasi-experimental design, we applied AI-empowered PBL to the experimental group and traditional PBL to the control group, comparing the differences between the two groups in computational thinking, data literacy, and creativity. The subjects were undergraduate students majoring in data science at a university in China, and the experiment lasted for one semester. Quantitative data included pre- and post-measurement scales (for computational thinking, data literacy, and innovation self-efficacy) and project outcome scores; qualitative data included interviews, code submissions, and analysis of AI interaction logs. Based on theories such as Vygotsky's Zone of Proximal Development, AI acted as a cognitive scaffold in the experimental group, supporting students in solving tasks beyond their independent capabilities through methods like code auto-completion and dialogue-based Q&A. The research results indicate that PBL integrated with AI can significantly enhance students' higher-order thinking and innovation ability: compared to the control group, the experimental group showed greater increases in problem-solving, critical thinking, and creative thinking abilities, with statistically significant differences; their project works also received higher evaluations in terms of complexity and novelty. Concurrently, qualitative evidence shows that students adeptly used AI tools for flexible exploration and immediate feedback, demonstrating higher engagement and autonomy. However, the study also found that improper use of AI may lead to over-reliance or the spread of bias; therefore, cultivating students' AI literacy is critically important. This research provides an empirical basis and practical framework for effectively integrating AI and PBL in data science education, offering implications for cultivating future-oriented innovative talents.