Infinite U (Building a Learning Engine) - Part 2.

Co-authored by Mickey McManus and Randy Swearer

Part 2 - A Thought Experiment About Generative Learning

Imagine a world where there are human learners and machine learners collaborating in high performance teams. Consider a shift from--the current mindset that fosters a "race to the bottom of the brainstem" and frames the conversation as a no-win situation where we'll all just need to give up on humans and put them all on the dole--to a more useful and powerful mindset that values the amazing potential of humankind. Where the products and environments we create and live and work within are themselves collaborators sensing and accelerating our capabilities based on the best of cognitive neuroscience and the best of computer science.

Maybe the first step would be a capability that would let us see into what we already know and do every day so that we could unearth our tacit knowledge and put value to the muscles we've already built. Alan Schoenfeld and other learning science researchers have outlined five critical dimensions (PDF) for both learners and teachers to explore when building new mindsets, skillsets, and toolsets. A key dimension is called "dynamic formative assessment" and is quite different than the classic "summative assessment" models that focus on teaching to the test. We wondered what would happen if we used sensors within a creativity tool to uncover how a learner demonstrates their existing skills through use--as well as to see if we could accelerate the mastery of new skills--through automation.

To give a hint at what that might look like a team of us at Autodesk Research have recently launched a free research plugin based on this research. It's called the "Command Map." We are working with product designers and makers to explore how they could dynamically discover and document their own learning "credit score" via the tool. Click below to see a short clip of the Command Map in action.

The Command Map also lets learners find learning resources and peer created content that can help them build skills to level up...

Learners can also discover which projects they learned the most from, how fast they learn and when they learned what tools or skill sets. This allows them to see which teams, projects, and organizations pushed them the most to learn and grow.

They can see how they compare to peers and discover areas they have deeper knowledge in as well as places others are honing their skills.

Team leaders can get guidance on which team members are ready to take on a challenge within the context of their entire team.

This is part of a larger initiative to build a lifelong learning engine that is itself a machine learning social network of smart credentials--watching the flow of teams and individuals as they struggle against challenges and grow in capabilities. The vision is that these new credentials would learn how to accelerate mastery the more students flowed through them and the more mentors and teachers interacted with learners.

Imagine if coursework continually adapted to the individual as he or she (or it) flowed through the curriculum. The learning units communicating to the last and next places the student could go. In computer science, curiously enough, this is known as a "deep learning."

As the system of machine and human learners evolves, new job roles that haven't even been given a name could emerge from the predictions.

"We don't know what you'll call this activity but it looks like there is a sharp rise in the combination of these skill sets, toolsets, and mindsets. Prepare now!"

Just like video games continually "test" the player's avatar by looking for evidence of learning during individual and group "quests" setting the gameplay to match the player, this new system could take a page from Adam Gazzaley's playbook and keep the learner right in the heart of that state of flow referred to as the "desirable difficulty" zone.

Teachers will shift to guides and peers on a common mission of discovery. Great ones will be more valuable than ever but with automation they’ll be able to reach and impact far more learners over their lifetime. Sometimes our teams of human and machine learners will discover powerful new ways to accelerate or deepen learning and create new smart modules that join the social network. At times the system will bubble up a new need and invite the market of module creators to build something new. Modules may even emerge at the intersections of credentials as the network inductively finds new patterns in the flow.

We’ll still need a way to certify inputs and outcomes and some sort of guiding body to make sure we don’t go off the deep end. The work of Credly around lower-case "c" credentials and the work of Credential Engine around building an exchange for capital "C" credentials may hint as the beginnings of this part of the ecosystem. Ultimately this will lead to a market of smart micro credentials, teachers, learners, mentors, environments and products all participating in a peer to peer acceleration of learning. Some credits will come from venerable institutions like ASU or MIT. Others may prove to be more effective and come from a lone module creator out in the hinterlands or a mentor that has found a way to accelerate mastery. Others may be discovered when physical spaces wake up and begin to notice actions in context that seem to improve learning "in the moment." It won’t be about a given framework or credit but about an emergent playlist of opportunities to grow a future “you.”

Its a Resource Allocation Problem.

Growing lifelong learners is all well and good, but the point of all of this is to in some ways address the massive resource disruptions that automation will create as jobs collapse and shift. Industry will be looking for agile, creative learners and will need to build a team to tackle whatever challenge emerges next. BCG recently released a set of reports about how leaders will "win in the 2020s." One of the key principles is that organizations will compete on the rate of learning. Today's fragmented credentialing systems--from capital "C" credentials like bachelor and masters degrees to lower case "c" credentials like a given person's ability to speak multiple languages or their mastery on a particular toolset--are stopping us from finding the best fit for an individuals next role.

What if the learning engine of the future could flag opportunities for you that were in your "proximal" learning zone? Something like, "While you may have incredible skills in designing factories for manufacturing you have almost everything you need to cross over into architecture, engineering and construction to help with the emerging field of off-site construction."

Imagine how people in charge of human and machine resources could discover the right team mates and grow them in new ways if the system had cross-sector liquidity and captured life-long learning...

Will team builders have a “What if?” tool to explore the potentiality of a given team as its born and grows? This person learned that because she had experiences within a certain place. That one has deep wisdom from a lifetime of diverse projects. This machine was trained by a team that pushed it to see unique insights. That person has the skills to forge the team into a whole new mind.

Infinite Playlists and the Golden Era

When we talked about "a playlist for a future you" a little earlier it may have conjured images of Netflix or Amazon Prime. That may not be a bad analogy. Consider the shift that these sorts of services brought to entertainment. In the previous era companies like Blockbuster served you any movie you'd like if it fit nicely into one of their six or seven genres and could fit into a little box that sat on a shelf. If you were interested in dramas or comedies, there was an aisle for each. Most of the titles were created by a monolithic system called, "Hollywood." There were plenty of choices for what pathway you'd take, as long as they were defined by the last century of movie going and a handful of respected movie making brands like Columbia and Paramount. They were in the entertainment business but it was very much shaped by the underlying "plumbing" of the movie industry. That is akin to picking a few pathways through school to get a degree that may have only been offered through a place like Harvard or Penn State.

Now consider someone like Netflix. It is in the entertainment business too but with the advent of streaming it shifted from a system-centric definition of entertainment towards a user-centered view and ultimately today we'd claim they have a learner-centered approach. Their playlist learns you and you learn about new kinds of entertainment over your entire lifetime.

They reverse engineered Hollywood and came up with a genome of entertainment that now has over 79,000 genres and a learning engine that continually pushes subscribers to explore right up to the edge of their comfort zone. Netflix and their kind create desirable difficulties and stretch their learners to grow every day.

Sound familiar? The playlist is learning you and one moment you're watching a Hollywood blockbuster and the next a small indie film and the next an entire TV series. That's like being able to learn a course at Harvard, then taking a workshop at a local maker lab, then trying out an Instructable tutorial online, and then taking a class from a guru on the other side of the world through one of the online academies. Except in the case of Netflix each activity helps grow your playlist and teaches the system about places you'd like to explore next. Effectively the system gets better the more you use it. In education, the credentials you can earn are fixed and how you learn doesn't get any better no matter how many people flow through a given course or educational experience. Worse, you don't get any credit for real world experiences or those that happen "out of band" or "on the job." All of our credentials are essentially dumb and dead, and our degrees are fixed and based on some industrial age notion of learning a static stock of knowledge we are supposed to deploy over a lifetime in jobs that rarely change.

Curiously enough Reed Hastings, the co-founder of Netflix has said that we have entered the golden era of entertainment. He and others like Amazon and HBO are now hunting for the best writers and production value they can get and paying top dollar for it. What if we could say the same thing for educators?

What if we could herald in the golden era of generative learning and what if the platform could help us discover the best of the best of teachers whether they are peers that sit next to you at your office or brilliant gems hidden on the other side of the world?

Next chapter? Could we build a Learning Futures Market?