Exploring the impact of ChatGPT on college students’ mathematical creativity: A quasi-experimental study

Abstract

Recent advancements in the use of ChatGPT in mathematics education have drawn significant attention from educators and researchers. However, the effect of ChatGPT on college students’ mathematical creativity remains inconclusive. Based on the hybrid human-AI regulation theory, this study addressed this gap through a quasi-experiment comparing an experimental group using ChatGPT to solve a mathematical creativity task with a control group that did not. The results showed that ChatGPT had a positive effect on college students’ mathematical creativity and enhanced their self-efficacy and interest in task resolution. Also, students using ChatGPT found the task easier and required less mental effort. Notably, ChatGPT may hurt the accuracy of metacognitive evaluation. Overall, the findings revealed that ChatGPT positively influences mathematical creativity, but it requires greater use of metacognitive skills. This study further refined the hybrid human-AI regulation theory by showing that improved performance and inaccurate metacognitive evaluation may coexist in human–ChatGPT co-creation. The findings offer valuable insights for guidance on using ChatGPT to enhance college students’ mathematical creativity.

Implications for practice or policy:

• Mathematics educators could use ChatGPT to enhance college students’ mathematical creativity, but this may impair their metacognitive evaluation.
• Policymakers should integrate ChatGPT into mathematics curriculum and develop courses in artificial intelligence literacy to enhance students’ metacognitive skills in the human–ChatGPT co-creation process.
• Educators could adopt a “humans first, ChatGPT second” approach to mathematics instruction, encouraging students to actively engage in metacognitive regulation during the human–ChatGPT co-creation process.