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

This portion was besides cross-posted connected the Civics of Technology blog. This portion besides has a followup station that you tin find astatine Three Questions connected Questions: On Asking, Knowing and Noticing

Two pieces crossed my provender recently, some astir Sal Khan and the AI tutoring gyration that wasn’t. The first was Matt Barnum’s reported portion successful Chalkbeat, wherever Khan himself acknowledged that Khanmigo, the AI chatbot tutor he launched 3 years agone pinch world-changing ambitions, was “a non-event” for astir students. “They conscionable didn’t usage it much,” Khan said. His ain Chief Learning Officer, Kristen DiCerbo, put it moreover much plainly: “So acold I americium not seeing the gyration successful education.”

The 2nd was Dan Meyer’s crisp obituary connected LinkedIn, titled “RIP Khanmigo & Edtech Industry Dreams of AI Tutors.” Meyer traced the full arc: the TED talk predictions, the philanthropic subsidies, the progressively fierce measurement Khanmigo inserted itself into the student acquisition (because students wouldn’t activity it retired voluntarily), and the steadily shrinking personification projections. His conclusion was blunt: if Khanmigo died pinch each advantage successful the world (early OpenAI access, Microsoft backing, authorities subsidies, Sal Khan’s Rolodex), what dream should the remainder of the edtech manufacture spot successful chatbot tutors?

These are important pieces, and I’d urge reference both. But reference them, I recovered myself reasoning astir a deeper question. Not whether the gyration grounded (it intelligibly did) but why it was ne'er going to activity successful the first place.

To explicate I person to spell backmost a spot successful time, backmost during my days astatine MSU, erstwhile I was retired location giving talks astir exertion integration and the captious domiciled played by the coach successful this full process. And further astir the value of students actively constructing representations of their understanding.

So successful these talks I utilized to show a clip from the Charlie Rose show (see below). It’s an question and reply wherever Khan describes really he prepares to thatch a caller topic. And it’s wonderful. Here’s Khan connected learning about, say, Napoleon and the French Revolution:

“I attack it from what my encephalon would for illustration to see… I for illustration to spot a scaffold, I for illustration to spot a map… what is the Holy Roman Empire, for illustration where, what is that now?”

He sounds Wikipedia first, “just to get the scaffold.” He draws timelines. He copies maps and pastes them onto his integer blackboard. And past he does thing genuinely important: he pushes past the aboveground until he hits the questions that textbooks skip. Here’s Khan connected the neuron:

“A biology book will show you okay the awesome goes crossed because there’s a myelin sheath and I’m for illustration yea but really does putting a small insubstantial astir a neuron, really does it make the awesome spell faster? And nary biology book will show you that answer.”

So what does he do? He ponders. He thinks it done by affinity (fiber optics, awesome amplification). And past he calls up friends who are biologists aliases communications engineers and asks: “Does this make sense?” Sometimes they corroborate his intuition. And sometimes, beautifully, they say: “You cognize what, we don’t know.” Khan’s consequence to that is perfect: “Why didn’t the book show maine that?”

I utilized to play this clip and past inquire the assemblage a elemental question: Look astatine everything Sal Khan does to study something. He sounds widely. He scaffolds. He draws. He questions. He calls friends. He argues. He makes connections crossed fields. He builds intuitive knowing from the crushed up. And past he builds thing to seizure each that he had learned to stock pinch others.

Now… really does Sal Khan want my kid to learn?

Watch a video.

The room ever sewage it immediately.

Because cipher successful that room would judge that for themselves. We each know, intuitively, that watching personification other explicate thing is not the aforesaid arsenic knowing it. We would ne'er settee for that arsenic learners. And yet, somehow, we judge it arsenic a solution for different people’s children. This is simply a type of a arena that I person written astir earlier: The reductive seduction of different people’s problems.

But there’s a 2nd furniture that I deliberation is moreover much important, and that neither the Chalkbeat portion nor Meyer’s critique rather names. Khan’s individual learning wasn’t conscionable active. It had a purpose. He was learning successful bid to make something. The video was his construction, his artifact, the point he was building. That’s what pulled him done the difficult parts, done the myelin sheath mobility and the calls to friends and the hours of immersion. He had a destination.

Students watching the video person nary specified destination. They’re receiving the merchandise of personification else’s learning process. And erstwhile Khanmigo came along, the gyration was… a chatbot to thief you person much efficiently. Still nary purpose. Still nary making. Still nary logic to push done difficulty. No wonderment DiCerbo reported seeing much “IDK IDK” than substantive engagement. No wonderment teachers astatine early-adopter schools recovered that students “didn’t really attraction for the bot.” Why would they? There was thing astatine liking for them.

This is wherever John Dewey, penning complete a period ago, becomes useful. Dewey based on that learning is built connected 4 earthy impulses: the impulse to inquire, to construct, to express, and to communicate. He saw these not arsenic skills to beryllium taught but arsenic drives already coming successful each learner, drives that acquisition should activity pinch alternatively than suppress.

Go backmost to the Charlie Rose clip and watch Khan done this lens. He is surviving each four. Inquire: the relentless “why” questions, the refusal to judge aboveground explanations, the “why didn’t the book show maine that?” Construct: the timelines, the maps, the blackboard drawings, the scaffolds he builds for himself. Express: the video itself, Khan giving shape to what he’s understood. Communicate: calling up buddies, testing his ideas against different minds, discovering together what is and isn’t known. And past building a practice of his learning, pinch his ain unsocial sound and style, and sharing it pinch the world. Inquiry, construction, connection and expression—all successful 1 go! Intermingled truthful good that it is difficult to show them apart.

All 4 impulses, firing beautifully.

And nary of them disposable to the student connected the different end.

Khan’s awesome error, I think, was not a nonaccomplishment of effort aliases sincerity. It was a nonaccomplishment of acquisition imagination. He knowledgeable the afloat richness of learning and past designed a strategy that offered students only the residue. He gave them the destination without the journey. And because the travel is wherever information lives, wherever intent lives, students rather reasonably declined the offer. First they declined the video (or rather, passively consumed it). Then they declined the chatbot. In some cases, the test was the same: cipher had fixed them a logic to care.

And this is what I deliberation the edtech world keeps getting wrong. The presumption is that if you tin present the correct content, successful the correct way, astatine the correct time, learning will follow. It won’t. Not without purpose. Not without the impulse to inquire, construct, express, and communicate. Not without, successful Dewey’s sense, the learner really doing something.

Teachers cognize this. It is, successful fact, a ample portion of what teachers do: return thing a student is not yet willing successful and create the conditions that make them interested. Not done tricks aliases gamification but done the creation of experiences that activate those Deweyan impulses. That is the work. And it is activity that nary video, and nary chatbot, has figured retired really to do.

I’ve written recently astir really evolution’s reply to an unpredictable world was not “more data” but play, and astir really children are optimized not for pattern-completion but for exploration. The relationship to the Khan communicative is direct. Khan Academy, and past Khanmigo, are some autocomplete strategies: 1 autocompletes explanation, the different autocompletes tutoring. Neither makes room for the exploration, the construction, the messy purposeful making that is wherever learning really happens.

Khan himself seems to person arrived astatine thing adjacent to this realization. “I deliberation our biggest lever is really investing successful the quality systems,” he told Barnum. That’s a singular condemnation from personification who has spent astir 2 decades trying to amended acquisition by routing astir the humans. Whether his benefactors successful the exertion manufacture will beryllium arsenic excited to put successful quality systems arsenic they person been successful package that tries to switch them… that remains to beryllium seen.

Endnote: I don’t usually bring TPACK into my blog posts. I mean, really overmuch weight tin a Venn sketch carry? But this mightiness beryllium the cleanest lawsuit I’ve ever seen. Khan has exertion knowledge successful spades. He intelligibly has heavy contented knowledge (that Charlie Rose clip is impervious enough). What he has ne'er had is pedagogical knowledge: an knowing of really group learn, what motivates them, what makes the quality betwixt personification who pushes done trouble and personification who types “IDK.” The circle marked P successful the Venn diagram. That is wherever the humans live. Content and Technology mean thing without that.