← Retour au blog
tech 23 August 2026

Why Sal Khan't: Learning by Making but Teaching by Telling

The failure of Sal Khan's AI tutor, Khanmigo, highlights a crucial truth about education: the difference between learning by doing and teaching by telling. Let's explore why it never truly took off.

Article inspired by the original source
Why Sal Khan't: On Learning by Making but Teaching by Telling ↗ punyamishra.com

Introduction

Education is an ever-evolving field where technological innovations often promise revolutions that never fully materialize. One of the most telling recent examples is Khanmigo, the AI tutoring initiative launched by Sal Khan with the promise to transform learning. Yet, three years later, Khanmigo is deemed a "non-event." Why did it fail? And what does it reveal about how we teach and learn?

The Story of Khanmigo

Khanmigo was launched with high ambitions, backed by partnerships with OpenAI and Microsoft, as well as government grants. Despite these advantages, student adoption was low. According to Kristen DiCerbo, Chief Learning Officer of Khan Academy, "I am not seeing the revolution in education yet." This raises questions about the effectiveness of chatbots as tutors.

Learning by Doing

The concept of learning by doing is supported by decades of educational research. John Dewey, an influential philosopher and educator, long advocated for experiential learning, where students actively construct their understanding through practical experiences. Studies show that students who engage in hands-on activities retain information better and develop critical skills.

Teaching by Telling

Teaching by telling remains a predominant method in many educational systems. This often involves a top-down approach where the teacher passively transmits knowledge. While this method is effective for certain subjects, it does not always foster active student engagement.

The AI Dilemma in Education

Chatbots like Khanmigo rely on teaching by telling, as they provide information and answers to student questions. However, they often lack the capability to engage students in experiential learning. Furthermore, recent studies show that human interaction plays a crucial role in learning, something AI cannot replace.

Use Case: When AI Works

Despite the challenges, AI has successful applications in education. For example, platforms like Duolingo use AI to personalize language learning paths, adapting exercises to the user's level. This personalization makes learning more interactive and engaging.

Conclusion

The failure of Khanmigo underscores the need to rethink how we integrate AI into education. Rather than simply replacing traditional methods with technologies, we must find ways to combine the best of both worlds: the active engagement of learning by doing and the efficiency of teaching through advanced technologies.

Let's discuss your project in 15 minutes.

AI in education Khanmigo learning by doing teaching by telling edtech failure
Deepthix newsletter · 100% AI · every Monday 8am

An AI agent reads tech for you.

Our AI agent scans ~200 sources per week and ships the best articles to your inbox Monday 8am. Free. One click to unsubscribe.

Visit the newsletter page →

Want to automate your operations?

Let's talk about your project in 15 minutes.

Book a call