How Khan Academy is designing AI for learning

In this episode of Education Futures, we speak with Kristen DiCerbo, Chief Learning Officer at Khan Academy, where she leads the teams responsible for content, product design, assessment, and learning science.

With a PhD in educational psychology, Kristen brings a rare perspective to the AI conversation: learning science first, technology second.

We explore how Khan Academy is building Khanmigo, its AI tutor and teaching assistant, and what it takes to design AI tools that support real learning rather than shortcuts.

In this conversation, we discuss:

• Why screen time is a poor proxy for learning
• How AI tutors can support practice and feedback at scale
• Why foundational knowledge still matters in the age of AI
• The growing concern around cognitive offloading and students delegating their thinking to machines
• How Khan Academy designed guardrails and safety mechanisms for AI used by children
• The tension between gamification, motivation, and real learning
• Why human relationships with teachers remain the strongest driver of learning
• What the school of the future could look like, combining technology with project-based learning

Kristen also shares how Khan Academy applies a risk-management approach to responsible AI, identifying potential harms early and designing safeguards directly into their systems. (more on this: https://blog.khanacademy.org/khan-academys-framework-for-responsible-ai-in-education/)

The takeaway: AI may transform education, but learning will always require effort, curiosity, and human guidance.

Read this to go further:

National Study in Top Journal Finds Khan Academy Learning Gains After Accounting for Key Unmeasured Factors : https://blog.khanacademy.org/national-study-in-top-journal-finds-khan-academy-learning-gains-after-accounting-for-key-unmeasured-factors/