Adaptive Learning in Elearning and Online Learning

Machine learning (ML) and artificial intelligence (AI) have been lately getting the attention of news outlets, particularly in the technology sector. Many articles have been published on ML and AI; they address the potential applications in our daily lives, many others speculate on how robots will take over jobs. It is true that people have lost their jobs to robots in some sectors but there are still plenty of areas where the limitations of technology prevent the use of them.

For example, technology advances in AI have made possible virtual tutors for medicine, child education, and some college tutoring services. Most of these systems rely on the use of a large sets of data to work. In these cases, the datasets are relatively easy to produce, that is, it won’t take a lot of time and resources to make.

In online teaching and elearning, adaptive learning is a buzzword for the application of AI to these activities. Adaptive learning refers to an intelligent system that tailors the educational experience per the needs of the student. As they progress through the course, the system determines the type of content, activities, and testing the student should learn from their performance in previous modules.

This means the database needed for such system should contain multiple types of the same content, for different levels of expertise. For example, the system could be structured with levels basic, intermediate, and advanced. The system determines the level of content the student needs, this is usually determined from a series of testing and activities at the start of the course. A student that is placed in the basic level will take a longer time to complete the course because they would cover more levels. A student in the expert level will finish the course faster.

Educators Can Use Adaptive Learning for Interventions

One of the difficulties adaptive learning faces is exactly the gathering of data needed to train the robot. Unlike other systems where you can mine the data to train your AI system, educational data is very hard to come by. Creating the amount of content that could produce such system will require many hours of development, which means a high upfront investment that not many can afford. Building an AI system is within the reach of any company but producing the content is very expensive.

But some large educational corporations (McGraw-Hill, Pearson, to name a few) have implemented adaptive learning in their educational LMS (Learning Management System). Their online textbooks have adaptive learning capabilities to help students achieve the level of mastery needed to succeed in the course. In general, adaptive learning vendors request datasets from the customers so they can train the AI system.

At this point, adaptive learning systems are used to assess a student’s level of mastery in a subject, but they are not used to run entire courses and this is because of a flaw in these systems: students find out how it works and try to game the system. This is a problem that has plagued adaptive learning platforms. For example, students try to find ways to defeat the system by pretending they have a lower level of mastery of the material, this way they face the easiest levels first.

Educators have found a great use for adaptive learning systems. They find it difficult to determine when a student needs intervention, unless the student asks for help (not likely). If they find out, it is too late to take any remediation steps. With adaptive learning, educators can determine very early in the course if a student is struggling with the material. This way they can design, in advance, remedial materials and devote resources to those students that need help.

In general, these systems tend to be expensive, not many educational institutions can afford it. The research is out there on the effectiveness of such systems in learning, it is not clear if individualized and personalized education is as effective as a course designed for all students. Although recent research has argued that students tend to have a better experience using these systems, whether they would have learned the same without adaptive learning is not clear.

Adpative Learning is a Complex System

The complexity of adaptive learning systems makes it a very difficult system to navigate. Students will need training to use it, they also need training on how to cover the material (that is, avoid gaming the system). Educators need to be trained on how to facilitate the course, but most of all, on how to read the data produced by students while covering the course.

But adaptive learning has other advantages worth mentioning. These systems tend to be more interactive than courses designed in a linear way, this means students tend to engage more with the material. Students are, in fact, in charge of their own learning, even though they might not know it. This modality is suited for self-directed learning approaches, where students can explore and learn at their own pace.

Whether adaptive learning will become the standard for the development of online and elearning courses is still an open matter. More research needs to demonstrate learning is improved using these systems. Finally, either the technology lowers its cost, or adaptive learning proves it is worth the investment of resources by educational institutions and corporate educational programs. In the meantime, most likely, your online teacher is a real person and not an AI robot, at least for now.