Moonshot Thinking and Personalized Learning

Moonshot Thinking

At SXSW this past spring, a Googler gave a presentation about what Google calls Moonshot Thinking, the kind of no-holds-barred, imaginative, off-the-wall, visionquest thinking that solves problems ten times more effectively than existing solutions do.  This kind of thinking was what spawned the Google[x] office, a home inside Google for pursuing radical ideas, like self-driving cars, email accounts with virtually unlimited space, and more.  (The “x” in Google [x] is for ten--as in improving what’s out there by a factor of ten.)  The man speaking at SXSW was Astro Teller, who oversees Google[x], and goes by the title “Captain of Moonshots.”

The session was an inspiration session, which means there wasn’t much in the line of content.  It meant to inspire with big ideas, but if you follow what’s going on in the tech world, most of it was familiar.  Despite these shortcomings, it did plant the expression “moonshot thinking” in my head, and I’m glad for that, because thinking beyond big is a good thing for us to do in education.  

Machine Thinking

I’ve been thinking about moonshot thinking recently, because I’ve been thinking about where we’re going with MOOCs, adaptive technologies, and the push for individualized, personalized education.  And just a little bit of thinking about what these innovations are accomplishing reveals that these “groundbreaking” innovations aren’t really all that sophisticated.  They’re making education more accessible--and that’s a great thing--but they’re not making it that much better.  Most MOOCs and online resources like Khan Academy are lectures with rewind and pause buttons.  That still makes them lectures. Many digital classroom tools, too, like clickers and other devices, have analog predecessors that are almost as or just as effective.  What we’re seeing now are incremental innovations, not radical innovations. 

But radical innovation--moonshot innovation, the kind of change that really would revolutionize learning opportunities--needn’t seem so far away.  And I was struck by a particular vision for how these technologies might converge, in a way that would truly change individualized learning, not only in its form of delivery, but also its accessibility and effectiveness.  I’ve been meaning to write my thoughts about this down, but the past six months have been a blur.  Reading Audrey Watters’ recent post about her forthcoming book, Teaching Machines, got me to sit down and spell this out.

Personalized Learning

Education media is obsessed with personalized learning right now.  But we must remember that when we talk about personalized learning, we’re also severely limiting what we mean when we talk about learning.  For the purposes of this post, learning has three levels:

• Level one: Knowledge acquisition (low-level learning: memorization, comprehension, etc.)
• Level two: Individual engagement of knowledge (high-level learning: synthesis, evaluation, creative work, etc.)
• Level three: Social engagement of knowledge (debate, discussion, authentic assessment, etc.)

The kind of individualized learning that’s all the rage right now is predominantly level one.  It’s learning material on one’s own in a mostly archaic, multiple-choice way.  There’s some video and some pictures, and for budding engineers there are opportunities to test code, but individualized online learning is overwhelmingly multiple choice, or not far from it.  The opportunities for true interactive learning are quite few and typically quite limited.  

The Future Technology We Already Have

But the opportunity for something akin to personalized human interaction, something akin to a personal tutor--a digital, conversational domain expert with human sensibilities and expression, complete with an interactive, multimedia projection... something akin to Jarvis from Iron Man, but perhaps with physical or projected form, something that could teach to us like a person to a person, but also through multiple sensory pathways--this kind of digital personal tutor needn’t be that far away.  

(Moonshot thinking = C3PO + R2D2 + Leia)

We already have conversations with our phones (Siri, powered by Wolfram Alpha), but there is software out there that is even better.  Every year there’s a competition called the Loebner Prize competition, which features the Turing Test, a test in which people essentially text chat with another party that is sometimes a person and sometimes a machine, and the goal is to guess whether you’re talking with a person or a machine.  The software is getting better and better at replicating human patterns of speech.  Writer and philosopher Brian Christian writes wonderfully about this in his book The Most Human Human.

We also already have something like 3D projection in the emerging digital eyewear industry.  Products like Google Glass and Recon Instruments’ Jet headset allow us to see into the space in front of and around us.  And, with commercial spatial capture tools like what we’re seeing with Microsoft’s Kinect (not to mention the more sophisticated eye tracking and facial recognition software that’s out there), we’re beginning to not only simulate 3D projection, but also interact with objects in three dimensional space.  

Couple this with a rich database of pedagogical know-how--not just rich databases of content, but of the kinds of questions that stimulate different kinds of thought, the kinds of visual images that prompt understanding, the kinds of relationships in space that promote different kinds of relationships--and the pieces of the puzzle begin to come together.

The Moonshot: One-On-One Machine Teaching

A machine with deep content and pedagogical knowledge that can speak conversationally is one that can teach in one of the most sophisticated ways we know: through Socratic questioning.  This machine--for which the foundational technology already exists; it only needs combining--could engage students in truly interactive learning: a conversation with an expert, a conversation in which a student asks questions or is challenged with questions or facts.  

Add to this projection capabilities and a database of images, and the machine could summon pictures, movies, or other media to aid in the lesson.  This would not be a difficult addition.

More complex, but still within reach with our current technology, are simulated projections in 3D, the kind of Minority Report and Iron Man three-dimensional representations that would allow students to interact with complex objects in space, to enable students to manipulate the shape of any molecule in high school science classes, or to walk through Egyptian pyramids in elementary school classes, or more.

Further, the kind of photorealistic digital representations of people that we’re seeing in the film and video game industries suggest that we could package all this into a projection of a realistic person.  A projection that could speak to us, challenge us with good questions, show us pictures, welcome us into and walk us through spaces, interact with us physically, present physical objects or diagrams to us in 3D space, and more.  In short, we would have a personal tutor and tour guide with the digital universe at her fingertips.  

And all this with our current technology, if we only take the time to mix the right parts together.

It’s a heady future--one which would help solve the primary problem with MOOCs and other online learning solutions: dropping out.  The completion rate for MOOCs is woefully low.  But, we know that ongoing learning requires motivation, and we know that motivation comes from autonomy, and we know that autonomy means the ability to ask our own questions, to follow the paths we find interesting.  Enabling this is a herculean task, but it’s not an impossible one.  And if it were developed, it would much more fully realize the democratization of personalized learning that we champion today.

Safeguarding What Is Human

It’s important to remember, though, that this isn’t the future of all education; it’s only a future of level one and level two learning: knowledge acquisition and personal engagement.  Our current efforts in education technology aim at moving this kind of work out of the classroom, and that’s a terrific thing.  As learning content--and, increasingly, skills--moves out of the classroom, what remains to be learned in the classroom will be character: ethics, socialization, attitudes, empathy, and more.  

The future province of school will be character work: the social interaction with content, how we engage the world together.  In this paradigm, we won’t need to teach content, and we likely won’t even need to be the “guide on the side” of the flipped classroom who helps students through difficult work.  The digital tutor will be available for that.  But, we will need to shepherd the impossibly complex work of student social mechanics.  We will need to bring students together in a social environment to learn and engage material together, problem-solving, collaborating, learning to divide and share.  

And we will need to do this in person.  The bulk of our socialization must be done with living and breathing people next to us, whose emotions we can sense, whose heat we can feel, whose electrical signals we can process.  There will be always be a place for school, or at least there ought to be.  The measured decline of empathy that has characterized recent generations is probably one of the greatest threats to our future.  We must remember that the end of education is and must always be about how we engage what we learn with the people around us.

Bringing the Future to the Present

At SXSW, I ended up leaving the Moonshot Thinking session early and reading the twitter feed and listening to the podcast later.  I might have learned more if I had stayed; it might have been my loss.  Whatever the case, the language of “moonshots” stuck, and again, I’m glad it did.  This post owes something to that talk.

Still, if a digital Astro Teller had been my digital tutor, I could have asked him/it questions to find out whether or not the content of the talk would have been relevant to my interests.  Or, perhaps, he/it, with the right pedagogical programming, could have asked me the questions to pull me in more effectively, to lead me more adaptively, responsively, visuo-spatially to where he wanted us to go.  

Or, he might have reminded me of the part of the talk that really did inspire me, the part I heard just a moment ago when I re-listened to the podcast, the part that asked, “What would you do tomorrow if you knew you wouldn’t fail?  And why aren’t you starting that today?”  I think I know at least one thing that I’d start.

We’re not far from it.  We only need to will, the inspiration, to move us towards it.  

(cover photo from the SXSWedu panel description)