Share this Post
In the previous article, I made the argument for an eCoach blending adaptive and affective learning with psychometric analytics. I suggested this engine could meet our current and future learning needs at both a trait and vocational level. But if such an eCoach existed, we’d need a content engine that could keep up, which is where generative content comes in. There are three broad generative approaches, human, contextual and algorithmic.
Human generated content
Unlike traditional digital learning experiences that are carefully planned and constructed over weeks or months in response to business needs, most human generated digital experiences are actually created spontaneously by people in response to an immediate need and motivated by a desire to share, to gain recognition or to provide help. These pieces of content are timely and relevant, can be pitched at various levels from novice to master and often evolve or surface and sink quickly.
The internet abounds with successful examples of this kind of content, from youtube, quora and wikipedia to the countless successful crowd based help or interest-based communities. There is emerging evidence to suggest approaches anchored by heutagogical and paragogical principles such as self knowledge, self direction and peer learning approaches, are also superior to traditional learning content development approaches, because they are better aligned with how humans learn socially.
We see this in social knowledge platforms like Bloomfire, Yammer and Confluence that seek to deliver a more natural, workflow integrated learning experience than the LMS. Unlike LMS’s that generally add social elements, these systems are fundamentally social, often built around an activity stream similar to Facebook. This approach disrupts the LMS paradigm, by integrating learning into your daily flow as an informal micro-interaction, by empowering the individual to be a teacher and by co-opting well established social media behaviors such as liking, tagging and following.
Some of these vendors are going further, using these platforms to disrupt hierarchy, divisional and organisational boundaries, by extending the social knowledge network to suppliers, partners and customers. What I especially like about this approach is its scaling of the master-apprentice model, so each person can have not one, but many mentors, interests and tribes.
Many of these platforms already offer low friction avenues for using the multimedia recording, editing and publishing power of smartphones to create and share content on the fly. But if you were to add other tools empowering anyone to easily produce richer learning content, this potential could be further harnessed. Of course many LMSs have content creation and review workflow tools built in, but in most cases these are traditional self-paced eLearning and quizzing tools, restricted to a handful of people within the organisation and subject to stifling governance procedures. In a way human generated content reflects a commitment to empowerment and trust that is simply absent in LMSs, that instead bestow content creation functionality on the privileged few, fearing chaos if the lunatics ran the asylum. But more about professionally produced content in my next article.
Contextually generated content
A subset of human generated content is context based content. This is content that arises from a specific activity such as a computer based process or an experience in the field. It’s not actually different from human generated content, but there is a distinct set of tools associated with its generation that I want to talk about.
Advances in image recognition are spurring the development of new kinds of context sensitive help tools like Leo and WalkMe. These clever tools provide real time workflow based learning on a device, by providing content specific to the screen, field or action you are on without having to program hooks into the software. Instead they associate the content to an image of the screen. This enables them to be deployed quickly across any computer based experience. Learners already create and share screen movies to help others, but they are divorced from their context. This new breed of tools, if made available to learners would enable them to create content directly within the context and make it available to anyone else who finds themselves in that same context.
A second category of contextual content creation tools that is currently underutilised are those that go beyond the power of smartphone to quickly generate, contextualise and socialise content (though traces is pretty cool), by offering specialised tools. The only one of these I know of is WildKnowldge, but I expect there are others. This tool offers a unique set of capabilities to consume and create rich, complex content. I think it represents a move towards an important part of future learning, the development of higher order thinking skills, wherever they are.
The third and final context based content category I want to share is on-demand virtual coaching marketplaces. There’s a few early movers out there, like Helpouts, Expertory, Expert360 and Everwise giving learners the capacity to get live help at the point of need, wherever you are. Unlike traditional live video support platforms, these marketplaces are not tied to a specific organisation, product or service, instead, you can get help or ongoing mentoring about anything, anywhere and in some cases record the session for future use. Like some of the social knowledge platforms previously mentioned, a great strength of these approaches is their ability to disrupt hierarchy, divisional and organisational boundaries.
Algorithmically generated content
The final theme I wish to touch on is algorithmically generated learning experiences. The benefits this approach offers overs traditional approaches are its ability to create in real time, its ability to personalise content based on data and its ability to self evolve.
The gaming industry is at the leading edge of this phenomenon with games like Left for Dead and Scribblenaughts using techniques such as vast object databases, dynamic game balancing, simple artificial intelligence engines and procedural generation to transfer some narrative control to the user, by generating scenarios and content on the fly based on the player’s choices.
Another approach used by games are possibility or problem spaces like the The Sims, Fallout, Mass Effect and Grand Theft Auto franchises. These games create vast open worlds and use brute force scripting to create massive databases of storylines and dialogue, once again enabling high levels of player control over the narrative. Additionally many of these games provide engines for players to create their own scenarios, landscapes and objects, creating a hybrid between user generated and computer generated content.
The difficulty with algorithmic approaches is that they struggle to provide an emergent narrative space that gives the player control over the storyline, without compromising the coherence and overall satisfaction that a more directed narrative delivers, but if the narrative is too directed, satisfaction once again takes a nose dive. But when you apply these mechanics to learning driven by an eCoach, it actually targets the sweet spot, because the learning pathway, anchored by the learners passions and goals provides a coherent narrative, while allowing the learning to unfold in a problem space and through interaction with other learners.
There are also hybrid approaches such as MMORPGs like World of Warcraft and EVE. These games are massive in scale, but everyone plays and creates in the same space, so the narrative is driven almost entirely by people, competing, collaborating, problem solving and creating, but still within the limits set by the game. An interesting sub-genre of these is augmented reality MMORPGs like Ingress that superimpose the problem space on reality.
If the eCoach engine described in the last article is the future, then we’ll need content generation approaches that will be able to keep up. What is certain is that the professionally developed, labour intensive and highly governed processes of the LMS will not keep up.
A value adding content space
So how do we bring these themes together to quickly generate content that surfaces precisely at the point of need and sinks when outdated? Content that combines the best people and computers have to offer? Content that is rich and nurturing of higher order skills? Content that can be generated on demand, sit in a narrative and be part of workflow context? Most importantly, how do we do all of this without breaking the bank?
I think the answer lies in shifting our thinking away from purely human content creation processes controlled by (and often living inside of) the LMS, away from tools used by just a handful of people to generate content, and towards collaborative human and computer value adding spaces.
Imagine a virtual learning space that records conversations you have with other learners, then with minimal support from the crowd, analyses them for relevance, meaning and context and converts them to a branched learning experience accessible to other learners via a computer generated character. Suddenly your nuanced tacit learning experience becomes available to others and over time, through crowd curation and augmentation, these characters become more sophisticated with branches to other characters, experiences and resources, creating an immensely rich and varied tapestry of learning pathways.
I could imagine this approach being applied to all kinds of learning interactions and content types, all of which could sit inside a simulation of the organisation itself, designed to enable exploration, experimentation and understanding, populated with real learners and simulated characters. Like the social knowledge platforms, these worlds could be open to customers, suppliers, channels and even the wider world.
Of course the assets within this world could be accessed as part of a workflow for just in time learning, but the real power of this sandpit would be its ability to support safe failure and drive innovation. The other massive benefit would be it’s ability to exist in a state of constant enrichment and renewal, based on creating resources from learning, rather than for learning by using humans and software as part of a collaborative content value chain.
This example is a bit far beyond our current capabilities, but not as far as you think. Companies like Smart3D capture, filtr8, Versu, Narrative Science, Automated Insights and the Michigan Tailoring System all offer relatively inexpensive platforms that automatically generate content, often in a narrative form and addressing various parts of the value chain including curated resources, 3D worlds, interactive characters, rich formative feedback and reports. Automated semantic analysis platforms such as Semlab , Ontea and Ontotext can parse signal from noise and make fairly accurate predictions about the meaning of and relationships between pieces of content. Also, large creative platform companies like adobe are making the move away from discrete products or bundles and towards cloud creative ecosystems, where the boundaries between products are dissolving.
So why am I so obsessed with rich complex learning experiences requiring higher order thinking skills and emotional intelligence? Because that’s where humans add the most value and because jobs primarily resting on procedures, processes and compliance will probably become computerised in the coming years. Also why am I so keen to see learning spring from an experimental virtual workplace sandpit? Well, Facebook has bet $2B on VR playing a key role in our lives, not just social or gaming. The medical and defence sectors have been using it for years and Gartner places at the beginning of the slope of enlightenment. Finally, if you contrast the success of Minecraft with Second Life, I would argue that Minecraft was so successful because it focused on creation, not just exploration.
Prescriptive, governed, instructional approaches that drive cost reduction and competitive advantage through better productivity and compliance are the sweet spot for LMss. But what organisations really need to do is nurture skills and activities that computers cannot yet master and link these not to productivity and compliance, but to innovation and intrapreneurship. eCoaches and value added user generated content chains can get organisations where they need to go and their inherently open, democratic, bottom up foundations make them fundamentally different to the LMS.
I want to finish with a real world example. We are designing a CPD solution anchored around self directed action learning cohort based projects for one of our clients. The projects driving these cohorts generate content from a community and from assessment activities that serve two additional purposes. The first is as the basis of projects for cohorts further along their learning journey. Their projects actually require them to build on the work of others. The second purpose is to rapidly (and where possible automatically) transform this evidence into digital learning resources for use by future cohorts. Granted this is not the vision of a virtual experimental space, automatically generated through human-computer collaboration I described above, but it does reflect the concept of value chains in which humans and software play a role.
It should also be noted that there will still be a role for professionally produced learning experiences in the future I’m proposing. It’s just that they will play a small role, not the dominant one. So in the next article, I’ll look at what these experiences might be like, drawing on insights from the entertainment, advertising, tech startup and digital marketing industries to argue the case for rethinking digital learning content.
David runs the “Kill the LMS” workshop, designed to disrupt your thinking about how humans learn, reflect on the limitations your LMS imposes upon the performance of your people and look at ideas and architectures to remove those limits. Click here to learn more about this workshop.
Share this Post