The Good, The Bad, and The Ugly in AI and Education

Introduction

Ben Cornell: Managing Director of Common Sense Growth and co-founder of EdTech Insiders
Dan Meyer: Director of Research at Amplify, former high school math teacher, and author of the newsletter Math Worlds

Background: Teach for America experience, mission to scale effective education
Emphasis: Shift conversation from tech to ed (education first)
Evolution: Agrarian to Industrial Revolution shift in education was about standardization and accessibility
Current Era: Requires new approaches due to dynamic world changes
Big Picture: Create adaptive learners, optimize educator capacity
Five Exciting AI Use Cases:
1. Assessment: Adaptive, real-time assessments that support educators and learners
2. Educator Efficiency: Enhanced lesson planning and grading
3. Learner Practice: AI aids in practice but not novel teaching
4. Data and Infrastructure: AI captures and organizes unstructured data
5. Parent-School Engagement: Bridging language divides, flow of information, and advice
Concerns: High-stakes areas like mental health and students with learning differences

Background: Lover of math and teaching, heavy responsibility of influencing students
Agreement: AI will be widely used, but skeptical of claims about revolutionizing learning
Concerns: Personalized learning's flawed assumptions
1. Schooling: It's about more than cognition; socialization and being human are key
2. Student Interest: Extended solo screen time vs. social learning
3. Classmates: Students don't see each other as liabilities
Challenges: Question of Gen AI's product-market fit and effectiveness in meeting educational needs
Critical Analyses: Whether personalized learning should even be pursued

Findings: Survey by Common Sense Media
- 51% of young people (14-22) have tried GenAI
- Only 4% use it daily
- Excitement vs. concerns about job loss and IP theft
Interpretation:
- Ben: Edtech aimed at adults; roll-out nuanced; importance of adult supervision in educational use
- Dan: Skepticism of integration needs; tech promises vs. actual utility; teachers need impactful tools
- Personalized Learning: Distinction between individualized and personalized; importance of multiple instructional modes

Strategies for Startup Success:
- Ignore investors; focus on professional development and implementation
- Align vertical structures in schools
- Recession market: schools want coherent, less-is-more tech
- Dan's View: Edtech operators' infatuation with product can overlook real educational needs
Challenges in EdTech Platforms:
- Effectiveness with only a small subset of students
- Need for critical reflection and alignment with educators
Role of Assessment:
- Ben: Change in assessment structures is pivotal for educational transformation
- Dan: Need for solid measurements in educational impact beyond academics, including social and emotional wellness

Overview: Changes in AI regulation at state level (Utah, Colorado, Connecticut, California)
Commonalities: Emphasis on user notice and transparency
Recommendations:
- Monitor regulatory activities
- Regular updates on AI output validity (false, misleading, etc.)

Age for AI Interaction: AI tools require adult-advised engagement for safety and appropriateness
Finding Collaborators:
- For Founders: Volunteering in classrooms to understand true educational needs; teachers as collaborators if needs are understood
- For K12 Environments: Funding should target problem spaces, not prematurely on AI; importance of pilot programs
Social and Emotional Wellness: Critical assessment and existing measures; integration of AI should support social unit

Exciting AI Tools:
- Ben: Ella (digital reading assistant), Magic School, Early Bird Learning (dyslexia assessment)
- Dan: TeachFX
Contact Info:
- Dan: [email protected]
- Ben: [email protected]
- Naomi (legal): [email protected]