ASU+GSV Summit Sessions Beyond the Calculator: Math's AI Revolution
Apr 18, 2025
Jon Woodard, leading innovation at Wolfram Research, shares insights on building AI math tutors. Sean Davenport, a Google Fellow and NYC Superintendent, explores AI's potential to enhance math learning for diverse student populations. Jiani Hou, an expert in online education from Think Academy, discusses adapting AI teaching methods from China for the U.S. Together, they delve into how AI can transform math education, tackle long-standing challenges, and foster deeper understanding, making math more engaging and accessible.
AI Snips
Chapters
Transcript
Episode notes
Wolfram's Unique AI Math Tutor
- Wolfram Research developed an AI math tutor that blends symbolic math with AI language models effectively.
- The platform allows customization at state, district, teacher, and caregiver levels for personalized learning.
AI Must Teach Logic, Not Answers
- Large language models (LLMs) struggle to teach math because teaching requires helping students internalize logic, not just delivering answers.
- Effective AI tutoring involves Socratic questioning and personalized feedback rather than simply giving solutions.
LLMs Need Tools for Complex Math
- LLMs are good at some math tasks but fall short on complex, multi-step problems due to their statistical word-prediction nature.
- Integrating specialized computational tools with LLMs improves their math reasoning and accuracy.