Improving Learning Outcomes in Mathematics Education through Directed Use of a Generative AI Socratic Tutor

Location

Space Coast

Start Date

29-5-2025 9:30 AM

End Date

29-5-2025 9:55 AM

Description

To improve the learning outcomes among community college math students, Hudson Valley Community College (HVCC) performed a preliminary study to explore the effectiveness of generative AI tools as aids to learning mathematics. Two sections (control and experimental) were selected from two math courses to undertake this study: Precalculus and Intermediate Algebra. The study employed Khanmigo, an AI Large Language Model (LLM) developed by Khan Academy in collaboration with OpenAI' s ChatGPT. HVCC was one of the first community colleges to deploy this tool in the classroom. Khanmigo can engage as a Socratic tutor for math students instead of just providing a final result. In the Socratic tutor approach, students ask plain language questions about a math problem, enter a learning conversation with the AI, and are guided toward the solution through a step-by-step iterative process rather than being handed the answer.

This document is currently not available here.

Share

COinS
 
May 29th, 9:30 AM May 29th, 9:55 AM

Improving Learning Outcomes in Mathematics Education through Directed Use of a Generative AI Socratic Tutor

Space Coast

To improve the learning outcomes among community college math students, Hudson Valley Community College (HVCC) performed a preliminary study to explore the effectiveness of generative AI tools as aids to learning mathematics. Two sections (control and experimental) were selected from two math courses to undertake this study: Precalculus and Intermediate Algebra. The study employed Khanmigo, an AI Large Language Model (LLM) developed by Khan Academy in collaboration with OpenAI' s ChatGPT. HVCC was one of the first community colleges to deploy this tool in the classroom. Khanmigo can engage as a Socratic tutor for math students instead of just providing a final result. In the Socratic tutor approach, students ask plain language questions about a math problem, enter a learning conversation with the AI, and are guided toward the solution through a step-by-step iterative process rather than being handed the answer.