Will The LMS Perish In The Age of Machine Learning?

Follow us on LinkedIn for our latest data and tips!

Will The LMS Perish In The Age of Machine Learning?

You probably have a Learning Management System (LMS).  It will present a set of courses and modules – usually e-learning, in which animated characters will talk you through policies and regulations and at the end you will complete a multiple choice quiz.  This quiz will have a pass mark.  If you score 80% you will be judged to be compliant for another year or so. 

In order to achieve this apparent competence, you may have needed to be cajoled.  Your line manager may have been asked to remind you and your team members at your regular meetings that everyone needs to complete the module.  Your HR Business Partner may have fired off a series of increasingly stern emails to ensure you get on to the LMS, launch the module, pass the test.

This may seem a pretty uninspiring view of the use and purpose of an LMS.  In my experience, it represents not only how many organizations use these platforms, but how most employees view them.  It is difficult to see how an enterprise system could be neither less relevant to building organizational capability nor more divorced from the business of learning.

The roots of the uninspiring use of most LMSs are found in their genesis.  The LMS started life as a way of ensuring compliance.  Using AICC standards (the AICC stands for Aircraft industry Computer Based Training Committee) the idea was to ensure that those working on aircraft had completed certain knowledge modules and achieved a passable level of understanding about how they can stop airplanes from falling out of the sky.  Like any airline passenger, I am immensely grateful somebody thought to license engineers and mechanics in this way.

Now I don’t know about your job, but most of the roles I perform – and the capabilities required to perform them – are not easily reduced to a series of ‘tick any which apply’ questions.  I am far from unique.  In most organizations with which I work, people have similarly complex roles.  This is not to say that flight engineers have simple jobs, but that the teach, teach, test model was never a comprehensive route to the competencies that they required and it is similarly partial in its ability to track capability in most other roles.

The LMS always had a limited function in facilitating and tracking learning in an organization.  But in many instances, it has been the only mechanism which has been used.  It has suffered from the classic technology problem, it has over-promised and under-delivered.

But there is another threat on the horizon in the world of technology enabled learning.  By definition, the LMS enables access to, and tracks completion of, relatively simple, rules based learning modules.  Primarily a way of presenting facts and testing (instant) recall, the LMS works in a world that is clear and unambiguous.  It exists in a work environment bound by right and wrong answers.

The LMS collects course completion statistics and pass marks.  The system tells the learning administrator whether mandatory courses have been completed and logs test scores. The LMS is a data gathering machine.  This is where the LMS is the possible predictor of its own obsolescence. Data drives artificial intelligence (AI) – the ability of computers to carry out so-called ‘smart’ tasks which previously required human intervention.  AI also drives machine learning – the ability of computers to scan patterns and make predictions or decisions about what comes next.  By gaining feedback on the quality of those decisions, the computer ‘learns’.

If we look at the contents of those e-learning modules best suited to the LMS environment, they are often focused on rules sets and decision trees.  E-Learning, as created to meet the narrow strictures of the LMS, is full of ‘if this, do that’ learning.  These competencies are precisely the tasks which can now be delegated to technology.  Earlier this year, a Japanese insurance company announced that the job of assessing insurance claims would in future be carried out via AI. The previous job holders were to be replaced as the prevalence of data and the automation of systems had reduced their role to simply following defined rules. AI doesn’t go home at the end of the day, doesn’t require a lunch break and doesn’t expect holidays. Once the machine has ‘learned’ what needs to done, the human judgment which used to be required is redundant.

To a certain extent, the rise and ubiquity of the LMS was the first step on this journey.  By reducing learning to rules based, multiple choice powered right and wrong answers, learning technologists created a dystopian world in which many jobs could be reduced to true or false.  If ‘competence’ equates to passing these simplistic tests, then expensive, unreliable humans naturally become an unaffordable luxury. The rise of the robots seems inevitable and we in Learning and Development have played our part in facilitating the takeover. Future employment patterns are threatened less by cheap overseas labor and more by increasing levels of automation.

Of course, it’s not all doom and gloom in the learning technology landscape.  Newer LMS platforms have sought to support different types of learning intervention.  Many facilitate peer-to-peer collaboration and curation of materials by those participating in learning activities. The reduction of course content to rules and tests is falling out of favor. 

But at their heart, most LMS’s are still built on the basis of course completions measured by simple tests.  This neither represents the complexity of most 21st Century jobs nor the capacity of digital platforms to support skill development and knowledge acquisition. 

The question facing the learning tech world is simple. Are we prepared to ditch rudimentary learning by rote and embrace the reality of our shared, complex future?  Or are we happy to be replaced by a robot?