The future of education

Episode Summary

Executive Summary: Susanna Loeb argues that education should be designed around the learner’s experience—especially relationships, personalization, and access to high-quality opportunities. The conversation covers tutoring’s proven post-pandemic value, the challenge of scaling effective practices across decentralized systems, and how AI could augment—not replace—teachers and tutors through faster feedback, differentiation, and better measurement.

Main Topics: Learner experience as the center of education (Priority: 5/5): Loeb frames education around what experiences help students thrive, emphasizing that relationships, engagement, and the right level of challenge matter more than formal structures alone. Pandemic-driven revival of tutoring (Priority: 5/5): COVID-19 exposed learning loss and disengagement, leading Loeb’s team to focus on intensive tutoring as the clearest evidence-based intervention for helping struggling students. Scaling proven practices in a decentralized system (Priority: 4/5): The discussion explains why spreading successful education reforms is hard in the U.S., where district-level decision-making, trust, and local variation complicate adoption. AI as an educational tool, not a replacement (Priority: 5/5): Loeb sees AI as a way to improve differentiation, logistical coordination, and support for teachers and students, but not as a substitute for human motivation and relationships. Research methods for fast-changing technology (Priority: 4/5): Because AI evolves quickly, Loeb argues for rapid A/B testing and close partnerships between researchers, schools, and companies to evaluate tools in real time. The future of education policy and practice (Priority: 3/5): The episode closes with a broad vision: flexible systems, strong partnerships, and a focus on delivering meaningful learning experiences to all students.

Key Arguments: Education research should focus on the lived experience of learners, because student relationships and engagement are what drive thriving. Intensive tutoring is one of the most effective interventions for students who are disengaged or behind, especially after the pandemic. Effective tutoring typically requires repeated sessions rather than one-off help, but the needed human workforce can be expanded through college work-study and teacher-prep programs. Scaling education reforms is difficult because the U.S. system is decentralized, so dissemination requires trusted messengers, proof points, and practical implementation supports. AI can help teachers with time-saving tasks, differentiation, and support for tutoring, but current evidence is still limited and mostly not causal for K-12 students. The best research strategy for AI is fast, iterative experimentation with anonymized real-world data and close school-industry collaboration. Human tutors remain essential for motivating disengaged students; AI is most promising as a complement that extends and supports human instruction.

Data Points: Years of podcast archive: 8 years - Russ Altman notes the show has been running for eight years and has become an archive of Stanford research. Randomized controlled trials on tutoring: 100 - Loeb says the broad finding comes from roughly 100 randomized controlled trials on tutoring. Typical tutoring frequency: About 3 times a week - Loeb describes the common effective tutoring dosage. Typical tutoring session length: 30 to 45 minutes - Loeb says effective tutoring is generally half an hour to 45 minutes per session. Education districts in the U.S.: More than 12,000 - Loeb cites the number of districts to explain why scaling reforms is difficult. Believable causal AI studies since ChatGPT: 20 - Loeb says only about 20 studies in the hub’s review met standards for believable causal evidence.

Pivotal Quotes: "for education, it's the experience of the learner that matters" — Russ Altman: Opening framing of the episode’s central theme "we actually know a lot about the experiences that help young people learn and thrive, and the real challenge and opportunity is figuring out how to make those experiences possible for every student" — Susanna Loeb: Future in a Minute takeaway on the main problem in education "we are working in a space where we know very little, where things are happening really quickly and that it's going to be important to set up ways of learning and adjusting quickly in real time" — Susanna Loeb: Her assessment of AI in education and why rapid evaluation is needed

Implications: Listeners should expect AI to augment education best when paired with humans, strong research, and flexible systems. The industry and schools need faster testing, better data sharing, and designs that preserve motivation and relationships.