AI in Teaching: Understanding Student Needs

Regardless of the level of teaching, we can list the components of teaching as the following: designing the class, understanding student needs, teaching effectively and responding to those needs, and measuring the performance. Among these components, designing the class and creating the content is done prior to class, and it is subject to updates throughout the class. Measuring the performance is done after delivery of a certain content and it also measures the effectiveness of the teaching method. The other two, namely understanding student needs and teaching effectively and responding to those needs require high interaction of students, and I think this is the most value adding part in teaching. However, as EdTech started dealing with more complicated problems, this is a question they started asking: can AI replace teachers? They might do teachers’ work partially, but I think the most challenging work for AI (robots) is to interact with students – especially in higher ed where the topics are more challenging, and students’ requirements might differ tremendously based on their background and abilities in different fields.

In a class section of 100 students, it is expectable that learning styles, abilities and backgrounds of students could be different. A teacher/instructor tries to handle this challenge by finding a way to deliver the content that is effective for the majority of the class. Sometimes it is possible to repeat some important points during lectures differently in order to increase understanding. Additionally, weekly office hours are held to help students grasp the topics better. Especially in distance learning context if we are able to profile students based on their different needs with AI and machine learning, we can tailor the content to be delivered specially for them and we can make their learning a lot more efficient. It is also possible that students’ attentions could be tracked (mouse focus and eye movements), and the content could even be tailored according to the student’s attention span. Another tool that is possibly very helpful is virtual assistant. Virtual assistant is an AI-powered teaching assistant who is available for students 24/7 – no need for appointments or waiting in the long lines for the assistance. IBM Watson has an AI-powered teaching platform that is mainly for K-12 which personalizes content, enhance students’ vocabulary and tutors students with AI- based tutoring. In Georgia Tech, Prof. Ashok Goel developed an AI-powered TA using IBM Watson, and the robot is called Jill Watson. Although it has a lot of shortcomings, Goel believes that it is about to get better with the future developments. Another AI-TA is Zcalled Beacon which is in use in Staffordshire University. One advantage of the AI-TAs is that as they get more data from individuals, they get “smarter” to respond students’ needs more.

As the recent developments show, as teachers and TAs, the help we can get from AI is a lot more than it used to be in the past. While some believe that AI Teachers can replace human teachers, some argue that it is impossible because these AI agents cannot inspire students and build personal connections with students. My thinking is that AI agents cannot replace humans in teaching but may help teachers and students tremendously in making the learning process more effective. And in the far future, with pre-recorded lectures and distance learning, I think they might replace teachers given that there is always a human mentor available in case of a need.