Artificial Intelligence or AI, as it is popularly known, is one of the most exciting areas the tech world has been abuzz about. Almost every industry and domain has been implementing or trying to identify use-cases that enable them to use AI or at least claim they have an AI component. However, many of these AI use-cases are developed just to create a buzz in the market, without adding any real value for its users or business. And, this is true for Learning Management Systems (LMSes) as well.
So, let us first start with defining what AI is. AI can be defined as ‘the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.’ This definition was correct no doubt, but in terms of expert systems that used to run on a defined ‘if-then’ AI logic. However, with Machine Learning (ML) becoming a key component for delivering Artificial Intelligence, these algorithms now have the capability to train and adjust with the changing circumstances, thereby continuously adapting and maintaining accuracy of output.
Therefore, I would define AI as ‘a computer program that is capable of constantly training itself, to allow the removal of human intervention for any decision making or provide a human like suggestion or advice or feedback with an ever increasing accuracy.’
One of the most popular examples of claimed AI component in the training industry is the ‘Recommendation Engine’. In many platforms, AI starts and ends with a recommendation engine. In fact, most of these engines truly work on basic algorithms that are popularity and administrator driven. And, in today’s day and age, these could at best be classified as ‘expert systems’. At the same time, they don’t really provide any true value and hence are the wrong use-cases to focus on.
Most of the training platforms are administrators/ L&D driven. A recommendation engine is something most would desire, but would still limit its use as their main focus is on getting training done. If an enterprise is truly determined for a Learning Experience Platform (LXP) approach, then this would be a top use-case that can aid self-learning with an option for users to choose and consume content of their choice.
For AI to deliver more (than an expert system can), it needs data to enable machine learning and run its algorithms. In training platforms, there is a limit to what data is being captured and processed for a user. Thus, this becomes a huge constraint for delivering quality AI output.
As some of the top AI experts suggest, the best way to identify a suitable use-case is to ensure that it fits well within the following parameters-
- Not too Specific
- Not too Easy
- Not too Unspecific
- Not Too Difficult
- Just the right fit
And, this pretty much stands true even for an LMS. For example, a use-case like ‘building more user engagement on LMS’ might be too unspecific and too difficult, while ‘recommending popular content to user’ might be too specific and too easy. Another important aspect of AI should be that it should not be backward-looking, but rather forward looking, which is to say that AI should not be utilized to analyze past data, but to give insights and projections of the future.
Now, that we understand these basic aspects of AI, let us try and identify some value-based AI use-cases on an LMS –
Automating Training Strategy for each employee
Addressing individual training needs in a large workforce is always going to be challenging. If an LMS’ AI algorithm can clearly identify a gap in an employee’s current skill/capability and builds a training strategy around that employee, I believe it will enable the L&D teams to avoid putting unnecessary time and efforts into running the platform in auto drive mode. These teams can then only focus on monitoring and helping the LMS increase its accuracy by providing clean and complete data. This would be in true sense an adaptive learning experience designed for a particular user.
This is in some ways an extension to the above use-case, where all you need to do is provide knowledge, information and assets to the LMS and it could guide and define content structure that helps build engaging content for your workforce. A good example of this could be seen in the iPhone’s Photos app, where an algorithm randomly (but in most cases smartly) uses a group of photographs and videos and builds a short video. In a fast moving world where your knowledge and understanding change at any moment, these real-time content development mechanism might truly change how useful content is being created and delivered to your workforce in real time.
Optimizing Instructor Lead Training (ILT) Costs
L&D teams spend heavily on ILT, especially where they need to manage internal and external trainers for different locations within a city, country or across the world. Most of the times, given the limitations of budget or alignment and availability of resources, certain programs need to be sacrificed. AI that can help optimize and efficiently utilize ILT resources and budget (maximizing delivery of necessary training) would certainly work wonders for L&D teams.
Today, many LMS vendors are projecting AI capabilities as part of their feature list and marketing campaigns. But, you need to ask yourself if an AI could truly address your problems or add any value to your L&D efforts.