Usable Math team publishes paper on ChatGPT AI for elementary school math learning.
Content
Usable Math project researchers Sai Gattupalli (a doctoral student in the College of Education), and faculty Robert W. Maloy and Sharon Edwards, have released a paper on their use of ChatGPT AI for elementary school math learning.
The paper "Comparing Teacher-Written and AI-Generated Math Problem Solving Strategies for Elementary School Students: Implications for Classroom Learning" compares teacher-written and AI-generated hints and strategies for solving fourth grade math word problems included in their Usable Math open online tutor.
Gattupalli, Maloy, and Edwards suggest that, when given the perspectives of a 10-year math teacher and four virtual coaches (Estella Explainer, Chef Math Bear, How-to Hound, and Visual Vicuna), ChatGPT-4 provides text-heavy, largely procedural problem solving strategies. Teacher-written hints offer more ways to differentiate learning for students through the use of:
- shorter, more child-accessible language;
- visual images;
- animations and GIFs;
- growth mindset-building statements that affirm children’s effort and persistence in solving math word problems.