Artificial Intelligence in Elearning

Artificial Intelligence in Elearning

Artificial Intelligence (AI) has been the subject of science fiction for a long time. Sci-fi authors envisioned wondrous and usually catastrophic scenarios for the age of widespread AI. No doubt, we are a long way from creating sentient, self-aware robots, but many things we take for granted nowadays would have been considered sci-fi only a few decades ago. We live in a world which is changing fast due to the exponentially accelerating evolution of science and technology. We must admit that AI is still in its infancy, but even so it already managed to creep into our everyday lives (parking software, smart sensors in cameras, personal assistants in smartphones etc.). The technology improves so fast that we can expect to see some incredible changes during our lifetimes, and AI is certainly going to be one of them. AI is going to change the landscape of many industries, and education will surely be one of them.


When you think about AI in education, first thing that comes to mind is probably a fully autonomous robotic teacher; it might come to that one day, but for now AI is most likely to take over more mundane tasks such as grading and content classification. As every teacher knows, grading is one of the most tedious tasks there is. Furthermore, teachers are expected to do it on their own time, and the same goes for lesson planning and content creation. These tasks are simply not included as man-hours; nevertheless they take up a considerable amount of time. It can take up to several days to mark all tests even for a lower grade class, simply because teachers have so many responsibilities; not to mention grading tests for a full college course. Even if you mark according to a key and have more than a couple of TA’s who would share the workload, it usually takes days to grade all the tests. With AI, this could be done almost instantly, and students could get a timely feedback instead of having to wait in suspense for days on end. An intelligent software can grade a test faster, more accurately and more objectively than any teacher could, almost completely removing the human error factor. AI grading might never be able to fully replace human assessment, but teachers would be able to fully automate grading of all multiple choice and fill-in-the-gap questions. Essay grading is still in the earliest steps of developments, but it will surely improve with time, and maybe someday even be able to simulate human grading.

Classification is another tedious and strenuous process. There is a considerable amount of content a teacher must go through in the process of lesson planning and material selection/creation. It is an enormous amount of information and it takes many a lot of time and subject knowledge to be able to do it effectively. An AI could do it more quickly and efficiently, and at a greater scale than any human ever could. Furthermore, it could find overlapping concepts and related fields in various subjects which would allow it to repurpose existing content or even create new materials. Going through all the materials in each subject to find related concepts would take a tremendous effort for a human teacher. An AI would be able to sift through it more accurately and at a fraction of the time it would normally take. This would greatly reduce the resources needed for content development and potentially improve instruction.

These functions would greatly reduce the workload of teachers and free up time for them to focus on aspects of their job which are uniquely human. Such as creative lesson planning, developing classroom activities and interacting and co-constructing knowledge with their students.

Adaptive Learning

The current studies on educational techniques indicate that one-on-one instruction from a human tutor leads to better understanding than classroom or online lessons. Why is that? As a tutor, one is capable of committing to the student; one is in a position to analyze the student and devise a learning path with his/hers particular needs, expectations and learning style in mind. In other words individualized learning gives best results. Unfortunately it is simply impossible to provide every student with a mentor who would be completely devoted to his/hers learning goals. Luckily, adaptive learning is the next best thing.

Adaptive learning programs are data-driven; they rely on big data analysis to track learning patterns and outcomes in order to develop better learning strategies and predict their results. Relevant data is extracted by models which are constructed by data scientists. However, the amount of data available today far exceeds the human capability to analyze it efficiently. According to a Harvard Business Review article, artificial intelligence can create thousands of models a week, as opposed to a couple of models produced by data scientists. With that in mind, it is essential to gather and analyze data about students’ learning styles, goals and behavior patterns in order to create an effective personalized learning experience. These systems respond to students’ needs, putting greater emphasis on topics they haven’t mastered. This kind of custom made education could allow students of different levels to build the same conceptual foundation and work together in the same classroom, and eventually bring those who are lagging up to speed. So there is no doubt that adaptive learning can be greatly improved by utilizing AI.

AI Tutoring

There will always be a role for teachers in education, but what that role is and what it entails may change due to new technology in the form of intelligent computing systems. As was already mentioned, AI can take over tasks like grading, content creation, but AI could even be adapted to many more aspects of teaching as well. AI systems could be programmed to provide answers to students’ questions or could even teach the most basic course materials. As a matter of fact, there was an AI posing as a teaching assistant at the Georgia Tech’s online masters of science in computer science program. “She” was named Jill and her role was mostly to answer student questions and provide information about the course. Some students suspected that there might be an AI around, but none of them were able to identify it. Nevertheless, there will never be a true replacement for a human teacher in education, but it is most likely that the role of the teacher will change. As a result of incorporating AI and elearning in general, the role of the teacher will shift from that of a lecturer to that of a facilitator. Teachers will provide their expertise and hands-on experience in order to guide students through the content and down their learning paths.

We live in a world that is constantly changing, and changing fast. AI has become a part of our everyday lives, and we didn’t even notice; probably because the AI systems that affect our world on a daily basis do not fit our concept of AI. Believe it or not, AI is already here. Google adapts results to users based on location, Amazon makes recommendations based on similar purchases, Siri and Cortana adapt to user’s needs and commands, and all web ads are geared toward each individual’s  interests and shopping preferences. These kinds of intelligent systems play a big role in how we navigate our virtual environments and live our lives online. They changed how we interact with information in our personal and professional lives, and are about to change how we find and use information for education. With newer, more integrated technologies, future generations will have an entirely different experience of doing research and looking up facts than the students of today. And AI will essentially enable them to learn at the speed of search.