Over the past few years, we’ve seen a lot of end of year blog predictions on existing problems to be solved in workplace learning for the coming year. Almost without exception these posts bring in AI as the "deus ex machina"; or as defined from Merriam-Webster: "as a plot device whereby a seemingly unsolvable problem in a story is suddenly and abruptly resolved by an unexpected and seemingly unlikely occurrence, typically so much as to seem contrived." So HOW will AI in practice solve our workplace learning issues? This week’s post focuses on one of the bigger struggles; learning content curation. Here are the five specific reasons AI will soon curate all our workplace learning content:
1. There is simply WAY TOO much content for anyone to directly manage!
From our own experience in most every industry there is a plethora of workplace content created and then forgotten. It’s all still there though. A few years ago it was accepted to create or buy some 3rd party learning content in the form of a course or micro learning module, assign it to your team and that was that. Today not only is the amount of relevant knowledge and content available to us exponentially greater than a few years ago, for many of us our jobs are getting more complicated making it difficult for any one person to keep track of all the useful materials to do our jobs. Also, as we’ve mentioned in other posts, the number of authoring mechanisms keeps growing. More people send emails, even texts in addition to the tidal wave of pdfs, MS Power Points, Word docs, Google docs and videos which contain key job skill knowledge.
2. We can now auto tag all content.
We’ve known for a while manual tagging content doesnt work. From above it takes too much time and the relative meaning of the tags can and usually does change over time. Learning companies like us and Docebo along with knowledge management companies like Bloomfire have successfully started to tame the vast amounts of untamable good, bad and ugly PDFs, MS Powerpoints, documents to align them by category. They do this through AI technology like LDA in an unsupervised fashion. This means any given body of content can by auto tagged without requiring human intervention.
3. AI based search is here.
Until fairly recently, searching for the right document was very difficult. It didn’t really matter if the document was in Sharepoint, OneDrive, Box, Dropbox - most any content repository. The missing element was applying natural language processing (NLP). For more on how we employ NLP in our search see our Oct 2018 blog: Feathercap now has AI driven search and curation. It’s the idea of being able to understand the semantic meaning of any question and matched to similarly understood meaning of a text segment within any accessible content. Then the correct page or citation with the answer can be pulled up. Further enhancements come from tracking all content viewed and by whom. This helps determine relevance and timeliness based on an employees time spent and role in the organization. This is the basis for assertions of “personalized” content delivery and adaptive learning. Coming soon one can easily imagine chat bots embedded anywhere with such NLP technology to enable a question to be asked either texted or verbally with the precise citation read back and a link to the citation available to view.
4. LMS, LXP, Knowledge Management and micro learning vendors will hone learning such as xAPI into all our content. We won’t need to manually remember which is the “good” content any longer.
This is something we and many others are starting to do; track every interaction of every piece of content. Currently many learning vendors offer authoring tools which traditionally allow authors to build animated, video and text slide multimedia experiences which are tracked both in time spent on each page as well as whether the user passed the minimum credentials of the learning content by assessment. We now can see how all content can be tracked using standards like xAPI. So not just tracking the first viewing or time spent to “pass” the content but everytime its viewed or experienced. This provides exponentially more data surrounding all of our content and further drives curation and search capabilities.
5. Because all searches and all interactions are tracked the amount of data will be hard to manually manage. AI systems will be better suited.
Most LMS (Learning Management System), LXP (Learning Experience Platforms), knowledge platforms and micro learning vendors tout their reporting capabilities for learning teams to visualize what skill gaps are present from their users and act accordingly. In reality, with the above huge amount of data that is generated it will be very difficult and time consuming for learning teams to do this directly. They will more likely relying on the above AI systems to make those recommendations.
See our AI and the Augmented Workforce primer on how a workforce and technology can effectively work on tasks together.