Algorithmic Generation of Content in eLearning
In March 2016, a Japanese AI wrote a complete novel and sent it for a literary award. The book, smartly titled as ‘The Day A Computer Writes a Novel’ did not end up winning an award at the Nikkei Hoshi Shinichi Literary Award ceremony, but came very close to it. Hitoshi Matsubara and his team at Future University Hakodate in Japan had gone through a process of selecting words and sentences and setting parameters before for construction before letting AI ‘write’ the novel. So, quite a bit of human effort had already gone into getting this literary piece see the light of day.
The idea of an algorithm generating content is not new. And before AIs actually start creating bestsellers on their own, it will pass through industries that can have immediate impact from algorithmic content such as eLearning.
Algorithmically generating instructions is what is already in vogue all around us. We hit upon it every time we play Angry Birds and everytime we stare at our screensaver. Some programmer had instructed the computer to mix and match different designs and that’s what plays on your screen. So, instructing a program to write content in a particular sequence is following the same pattern that we have mentioned above. In fact, in our interactions with earlier site like ‘Summify’ we have come across similar activities. An algorithm sums up news content and presents the gist of it.
When it comes to eLearning, the LMS itself is designed for an easy mix and match of course modules so that customized learning can be provided as per the need of the day. This entire task of mixing and matching, tracking a learner, customizing module arrangements and creating content bundles can be automated with algorithms with an courseware designed pre-packaging all related tags and information in the algorithm beforehand. Here’s a possible scenario – Imagine John as a learner who has recently been shifted to a new job role in sales. He is supposed to understand products for the packaging industry. But John has been selling something else till date, so his competency is low in this. He needs to pick up. He can do a self-check of his competency level, which can inform the LMS how much John needs to pick up. But there’s no ready course in the LMS for John. So, the LMS digs into the micro modules of each course (where each micro module is a shareable content object) and then starts arranging them together to address John’s competency requirements. It also digs into the question bank and creates a set of pattern questions that can evaluate John at the same.
So, while John finishes his self-test, goes for a coffee and comes back, his course is ready for him to start.
Normally, this would have been scary scenario for courseware developers, however, the brighter side of the story is that you can concentrate on the high end work of imagining the scenario, creating the branched nuances and creating an engaging experience for the learner.