The near-future for learning is exciting.
Even more importantly, the technologically-enhanced future of learning will be more equitable, bringing the right content to anyone, anytime, anywhere, regardless of ethnicity, politics or background.
Web technology feeds learning technology
Learning pedagogy, technology and practice inevitably draw on (but tend to lag behind) the developments of the web, the world’s main stage for advancement and innovation.
Ten years ago, the authors of “E-Learning 3.0: Anyone, Anywhere, Anytime and AI” made this connection. They point out that the first wave of e-learning content supply and the LMS is really a Web 1.0 technology, bringing instant scalability but being read- and admin-oriented. LXPs (and progressive LMSs) are more Web 2.0, being more read-write and social.
Web 3.0 and web3 will bring the benefits of artificial intelligence to everyone.
Already, today, natural language — and therefore almost all learning content that exists — is machine-readable for specific, defined usage. Content intelligence is one live example of that. Any learning can now be read by machines and classified usefully in order to procure, curate and find content better. The semantic web vision for learning content has been realized (though it’s broadly considered to have failed for the web in general). Note that with this application of AI, the data is the content and the content is vast; the data gap was filled first and the AI solution followed.
Tomorrow could be extraordinary. Many of the crowning jewels of Web 3.0 and web3 have been designed to be open source, user-friendly and ship with APIs, such as OpenAI’s GPT3, which generates natural language to an expert human level, seemingly at will. This means that the time between the launch of cutting-edge technology and it reaching corporate learning will decrease substantially. Learning might finally advance from the back seat to a board seat. There is already a growing list of GPT3 content creation tools that will impact creators, publishers, academic and corporate education materials as well as the design process.
We’re less than five years from this. The technology is here already. What’s missing is the data.
Good data
Before indulging ourselves in the promises of state-of-the-art and near-future technology, we, in the learning and development industry, need to get our data in order. The enormous, interconnected web records and shares just about every tap, swipe, click, search, post and tweet. This underpins all the functionality, sophistication, growth and possibility of the internet.
Learning data still has a way to go.
Each year, the average knowledge worker reads 68 articles, watches 45 educational videos and listens to 14 podcasts to increase their confidence and competence at work. They watch far more YouTube videos and scrolls through far more social media feeds. The reality is that in learning, we have less activity to track than other more consumable media. Worse still, what learning there is, happens in different places. There is no single repository for this data. Indeed, we’ve had to concoct these very statistics, precisely because they don’t exist anywhere!
Some corporate learning systems — LMS, LRS, LXP — are starting to fill the data void. But these systems don’t capture most actual learning, nor do they stay when that employee moves on to another organization and furthermore, they are restricted by regulation, such as GDPR. So, users repeatedly find themselves back at square one, and this in itself inhibits learning enthusiasm and activity.
Furthermore, data is ringfenced for each organization. Private networks — even at very large companies — are too small for AI to reliably spot patterns, predict and recommend. Within a single company, we barely feel any of the benefits of economies of scale or network effects or the pooling of data for the greater good.
The logging and sharing of data is also formally restricted by concerns about privacy, personal identifiable information and associated regulations.
So, owing to a variety of entrenched factors, we have little to begin with and are unable to share what we have fruitfully.
How to get better data
Here are some ideas, drawn from Web 3.0 and web3 developments, to get the learning data flywheel spinning. We put these to adventurous, ambitious and frustrated colleagues in L&D, HR and beyond to stimulate ideas and action, to open up possibilities for the future.
Tokenization is the process of replacing potentially sensitive information with less sensitive information — in other words, anonymizing it — while preserving its business utility. So, rather than “this person learnt this thing at this time and date,” we might do almost as well with the less specific, “a person learnt this kind of thing in this period.”
Open-source experimentation (perhaps a more realistic aspiration than open-source data or open-source algorithms) is another method of reaping many of the rewards without much of the risk. Four specific suggestions from that article include: limiting the audience, limited disclosure, individual opt-in and summarization.
Decentralization. An open, decentralized learning ledger (a blockchain for learning) might solve both the ring-fencing (companies tend not to share data) and permanence (data is lost when I change employers) issue. “Open” implies a certain level of sharing. “Decentralized” means that no single entity can bring an end to the arrangement.
While it may be difficult to validate that actual human learning has happened, it may be enough that individuals register the content they think is useful, now and in the future. Such a register also brings a positive short-feedback loop (you add an item, your count goes up by one, your professional status is enhanced). There is a groundswell of blockchain-based learning ideas.
All of these ideas for generating more useful, shareable, equitable data require a shift in attitude. We will need more inter-company collaboration, even between erstwhile competitors. A change in mindset leads to improvements in data which leads to a learning nirvana — enhanced and equitable. The “E” in DEI should also mean learning equity (i.e. fair, impartial education for all) and open learning and learner data is a part of the solution. If you’re reading this, you’re probably one of the people we need on-side.
