As an e-learning course designer or as the person responsible for the LMS in your organization, the hardest part of your job is content creation. It’s a repetitive and time-consuming task that requires a deep understanding of the subject matter combined with continuous assessment of the needs of your students.
You might wonder is all or part of this task can be automated. You might even have heard about “algorithmically (or computer) generated content”, and set your hopes high. In this post we’ll have a look into “algorithmically generated content”, explaining what it is all about, how it fits with e-learning and how mature the technology currently is.
Algo-what?
An algorithm is a just another word for “computing instructions”. Algorithms are the constituent parts of computer software. You can think of them as “recipes” for achieving a certain outcome. The steps to calculate a square root? That’s an algorithm. The steps to decode and play an mp3 file? That’s an algorithm. How the projectiles move in Angry Birds? Yep, that’s too described in an algorithm.
Inventing algorithms and translating them so the computer can execute them is what programmers do. And “algorithmically generated content”, perplexing as it may initially sound, describes the simple idea of instructing the computer to generate content.
Algorithmically generated content is nothing new.
It exists, in various forms, for over half a century, while the related research goes even further back. Consider a typical screensaver, the kind that displays colorful graphics on your monitor. That’s “Algorithmically generated content”, right there. The programmer didn’t draw each individual picture in your screensaver. It just gave instructions to the computer on how to combine simple drawings and effects, and let it create thousands of pictures on its own.
The idea behind algorithmically generated content for e-learning is similar (though the algorithms are much more complex). We feed the computer with knowledge about the subject matter, and give it instructions on how to combine it to generate new material.
Yes, but is it any good?
Well, kind of. There are several algorithms (or “computer programs” if you prefer) for algorithmic content creation. Some of them you might have stumbled upon, in services like (now closed) Summify, which provided computer generated summaries of news articles, or in websites such as the Big Ten Network, that feature sports stories written by computers.
Existing examples of algorithmic content creation are somewhat crud (like the aforementioned automated summaries), or are based on rigid rules and data and produce formulaic results (like Big Ten Network’s sports stories, or computer generated weather forecasts). So don’t expect to be able to tell the computer “write me a 20-lesson course on astronomy” any time soon. Maybe that’s for the better though — you wouldn’t want computers to put you out of business as an e-learning course creator, would you?
So, is algorithmic content creation of any use to me now?
Yes, it is. You might still have to write the better part of your course material, but there are several ways algorithmic content creation can help you. Even today, capable LMS systems, can automatically create quizzes and exercises drawing from a pool of existing material and a set of simple rules. That’s a basic form of algorithmic content creation that is nevertheless a huge time saver.
Perhaps the more immediate success will be that of a hybrid form of human assisted algorithmic content creation, in which the content creator writes, tags and categorizes his course material, and the LMS engine uses it to create different courses for different groups and students with different skill levels and interests (adding supplementary material for things that the students have difficulty with, using test results to reinforce specific topics, tailoring a course’s progress to user interests, skipping some chapters, etc).
This kind of computer assisted content creation will be a much bigger deal in the future. And it’s perhaps this synergy, of human and machine, building on the strengths of both, that will define tomorrow’s e-learning courses, helping drive down their cost, and enabling a much better customization of the material to the individual student.
