What kinds of struggle are essential to learning and what kinds of drudgery should we be happy to hand off to AI?
Leah Belsky, who leads education at OpenAI, joins My Robot Teacher to talk about one of the hardest questions facing educators in the age of AI: where to draw the line between productive friction and work that technology can usefully remove. Drawing on her work at OpenAI, her experience at Coursera, and her own life as a parent, Belsky makes the case for preserving the forms of difficulty that actually build understanding while using AI to make learning more accessible, personalized, and manageable.
The conversation then widens beyond the individual student to explore what AI might enable at the level of classrooms and groups: collective intelligence, project-based learning, new forms of critique and collaboration, and learning environments in which students become builders rather than spectators. They also discuss age-appropriate guardrails, why secondary education may need to become more interdisciplinary and maker-oriented, what the United States might learn from Estonia’s approach to technology and teacher agency, and why educators who are worried about AI need to engage with it rather than retreat from the conversation.
Also in this episode:
Where the line falls between productive struggle and educational drudgery
Columbia’s CAiSEY project and using AI to make case-based learning more interactive
OpenAI’s new tools for study planning, flashcards, quizzes, and other student workflows
Why Belsky would not put general-purpose chatbots in the hands of young children
The case for teaching high school students to build and create with AI
Moving the unit of AI design from the individual learner to the classroom and collective intelligence
How AI can make project-based and collaborative learning possible at much greater scale
Using AI as a “seminar secretary” to understand how groups think and learn together
Why AI may blur traditional professional roles and push education toward greater interdisciplinarity
What Estonia’s approach to privacy, teacher empowerment, and student responsibility might teach the United States
Why treating AI only as a productivity or cost-cutting tool could undermine education
Why educators need firsthand experience with current AI systems before they can meaningfully shape their use
CHAPTERS
0:00-6:40 - Chapter 1 Introduction: What Should We Delegate to AI?
6:40-10:23 - Chapter 2 Leah Belsky’s Journey to OpenAI and the Promise of Accessibility
10:23-16:58 - Chapter 3 Drudgery vs. Productive Struggle
16:59-20:35 - Chapter 4 Age-Appropriate AI and Rethinking Secondary Education
20:35-29:59 - Chapter 5 From Individual Agency to Collective Intelligence
30:00-40:05 - Chapter 6 Estonia, the United States, and Two Cultures of Educational Technology
40:07-44:28 - Chapter 7 Risk, Equity, and Why Educators Have to Engage
44:29-50:34 - Chapter 8 Reflections: Codex, Canvas, and the Faculty AI Epiphany
My Robot Teacher is hosted by Sarah Senk and Taiyo Inoue, produced by editaudio, and funded by the California Education Learning Lab, incentivizing innovation in teaching and learning in California’s public higher education.
📄 Full transcripts available on Substack: https://calearninglab.substack.com/s/my-robot-teacher-podcast
🌐 More about the show: https://www.myrobotteacher.ai
📨 Email us! myrobotteacherpod@gmail.com
🔔 Subscribe for extras: https://www.youtube.com/@myrobotteacher
