PME-NA

Creating, Comparing, and Sorting Textbook, Teacher, and AI-Generated Word Problems for Third Graders

• In Event: Tasks in Elementary Years

Description

This study explored third graders' ability to distinguish between AI-generated, teacher-created, and textbook word problems while examining what makes AI-generated problems comparable to traditional sources. We worked with eight third-grade students from a high-ability classroom, having them sort and explain their reasoning about 33-word problems (14 AI-generated, 14 teacher-created, and five from textbooks). Students correctly identified the source of problems only 36% of the time. Their reasoning revealed interesting assumptions about how AI and teachers generate problems, such as expecting teachers to use smaller numbers or certain words like "really." Our results suggest that while AI can create problems that many students cannot distinguish from traditional sources, careful attention to prompting and language is crucial.

Authors

• Yi Zhu, Purdue University

• Luke Anthony Braun, Purdue University

• Sezai Kocabas, Virginia Commonwealth University

• Laura Bofferding, Purdue University