Skip to content

Interview with a Professor: What AI’s Disruption in Education Means for Students

The holy grail is personalized learning. Each student will be provided learning opportunities and examples that will allow them to learn at their speed and preference.

Ray Schroeder, Senior Fellow at the University Professional and Continuing Education Association and Professor Emeritus and Associate Vice Chancellor at the University of Illinois Springfield.

Artificial intelligence technologies have already shown promise in supporting educational markets. As online and tech-driven education solutions come under the spotlight in the fallout of the pandemic, governments and private companies globally are increasing investment in AI to spur the growth of the tech in support of future educational needs.

AI can also help strengthen students to make return-on-investment (ROI)-based educational decisions as well as enable them to pursue lifelong learning. We talked to a professor with deep expertise on the topic to learn about what AI’s growing disruption in education may mean for students and the classroom.

Meet the AI Expert: Ray Schroeder

Ashok K. Goel

Ray Schroeder, Senior Fellow at the University Professional and Continuing Education Association and Professor Emeritus and Associate Vice Chancellor at the University of Illinois Springfield

Ray Schroeder is a senior fellow of the University Professional and Continuing Education Association (UPCEA) and professor emeritus and associate vice chancellor at the University of Illinois Springfield (UIS). He launched UIS’ online learning program in 1997, founded the university’s Center for Online Learning Research and Service, and became associate vice chancellor for online learning in 2013, a position he continues to hold.

Schroeder regularly publishes articles and book chapters, and presents nationally on emerging topics in online and technology-enhanced learning. He is the recipient of numerous national awards and citations for individual excellence and leadership from various associations and entities, including the Sloan Consortium, the U.S. Distance Learning Association, the American Journal of Distance Education, the Illinois Council for Continuing Higher Education, the University of Wisconsin and the University of Illinois.

Q&A with Ray Schroeder Can you define and explain the relationship between artificial intelligence, machine learning, deep learning, and quantum computing?

Professor Schroeder: First, let’s begin with artificial intelligence. And when we talk about AI we look at several versions of artificial intelligence. We can see that AI uses advanced networking as well as computing with high performance computers, and with that we can perform machine learning and deep learning. We use algorithms and realistic, supervised, and unsupervised learning. So those topics are ones that generally I think the public ought to be aware of.

Machine learning is really the application of AI where it can learn on its own and improve from what it has learned. So, you establish a program that allows the computer to seek information on its own and to assimilate it into an enhanced program or database.

Deep learning is using machine learning functions in that similar way, but it becomes progressively better and better. Not only does it gather information, but it refines its own program. And that deep learning can continue to improve over time. Depending upon whether it’s supervised or unsupervised, the programmer will have a hand in reinforcing and saying yes, go this way or no, don’t go that way. This is happening now and has been for the last several years.

When we talk about quantum computing, that’s like a super supercomputer using a different kind of technology than we use in our current computers. We’re comfortable storing information as bits. For example, if someone writes a report and wants to save it digitally or send a picture over the Internet, you can see how many bytes comprise that file. Bits are binary digits, 1s and 0s. The word “bit” is a combination of “bi” from binary and the “t” from digits. Eight bits make a byte. So, bytes are composed of groups of eight bits.

The difference in quantum computing is that it uses qubits—quantum bits—and these are different in that they use subatomic physics, which is a bit strange to us. The qubits can store information as a 1 or a 0, or a 1 and a 0, and while that doesn’t seem natural, it is done and it enormously speeds up the ability of a computer to process information. So broadly, quantum computing enhances the artificial intelligence programming as well as the execution of programs. Quantum computers are the next great, super supercomputer that accelerates artificial intelligence.

AI is most important for all of us where there are huge data sets. Otherwise, you can get by fine with basic filing systems. But what AI can do particularly well, is take a large amount of data, sort it, and make sense out of it for you. And using supercomputers you can do that much more rapidly. Where do you see the current applications of these technologies in education and room for adaptation?

Professor Schroeder: There are some really exciting things going on with artificial intelligence. For example, there are some online musicians that are actually artificial intelligence programs. AI can write music, use a voice, and even perform their music without intervention. And now this programming is selling its own music and depending upon responses, it can follow genres, and adjust and improve its work. So that’s useful, for example, if you’re studying music history; if you’re studying music in general, there’s much that can be done with AI in that area.

Certainly, AI is a tool for our assistance. And we’re going to see far more of that in education. For faculty, it can provide all kinds of services, even including grading and that has been around for quite a while. One of the things that I’m doing now, which is new to me, is fourth-grade instruction because I supervise or assist my nine-year-old grandson as he is taking the remote learning and we’re using the tools.

Among the tools is adaptive learning, which is AI-driven, so that when he takes a quiz after a lesson, depending upon the right or wrong answer on the quiz, it advances him. In some areas like math, he’s up to a sixth-grade level despite being a fourth-grader early in the semester. So AI will advance him at his speed. And if he’s having difficulty, it will continue to provide different examples and support him with learning materials that are slower speed so that we can thoroughly learn those materials. And ultimately, I suppose that’s where we hope to go with education.

The holy grail is personalized learning. Each student will be provided learning opportunities and examples that will allow them to learn at their speed and preference.

I taught for 45 years and I always stood at the middle of the class because you didn’t want to get things confused if they were on the slower side. You didn’t want them to get bored if they’re on the upper side, so you had to add assignments.

What artificial intelligence allows us to do, using what we call adaptive learning, is to begin with our learning outcomes, know what we want to get at the end, and then people set the students through learning with the AI program, placing them along the way and changing the timing and the types of examples that it gives to the students. That’s a hugely important tool in education.

There is also use of AI as a faculty assistant specialized for education that can remind you to do XYZ, and grade papers and so on. But what I think many faculty are not thinking about yet is that students have these, too. Some students have actually written their own advanced assistants that can write term papers—assistants that can go out and do research and return it to them.

For example, you could say, “I need a 10-page paper on comparative analysis of the development of COVID-19 in Europe, Asia and America, and I need at least 12 annotated sources for this in APA Style. Please include a conclusion that compares how Asia and America responded and which had the best results. I need that in 15 minutes, and I’d like you to send it to the printer, two copies. Also, send it to me in an email and I’ll send it to my professor.”

So that changes the whole process. And in fact, a high percentage of news reports, from the Washington Post, Bloomberg, World Report and Associated Press are written by AI—not reporters. Bloomberg says one-third of all their reports are written by artificial intelligence. Some of those are fairly easy like sports, where the computer has all the terminology: “He hit a three-point shot” or “They did a double play.” All of the terminology specific to the sport can be collected quickly, passes through the editor, and goes out to readers. Clearly, potential applications for AI are extensive. What does this mean for students’ data and privacy? And what other questions are being discussed in regards to AI’s impact on learning?

Professor Schroeder: Yes, that’s a good question and there’s definitely discussion on that. Maintaining the privacy of students is important. Right now we’re doing that anyway because we’re using learning management systems. We’re recording electronically at the universities and have to focus on privacy for those uses. But still, it’s an important part of the process.

I would say that adaptive learning is key to enable effective learning for students at all levels. And it can really help in bringing a greater diversity of students and help to build the bridge for students to universities that have been underprepared or underserved by their K-12 experiences. So in that regard, AI is enormously useful.

Can it teach? One of the things that I [like to] share is Ashok Goel’s work at Georgia Tech. I’ve had the good fortune to have lunch with him a couple of times. He teaches artificial intelligence and master’s in computer science, and he developed the Jill Watson AI assistant with a Watson computer. When he was teaching a program with a class of 300 students, he was given six teaching assistants (TAs) to help answer students’ questions and keep them moving along.

He also added a seventh computer AI TA that he called Jill Watson. In the beginning, she was very awkward but by the end of the class, some of the students nominated her as TA of the year. And when they surveyed the class, not one student knew or said that they knew that she was really a computer program. When conversing with students, she would throw in anecdotes about the weather and sports and all kinds of things to be conversational.

So what Goel has been doing over the last five years is refining this program so you don’t need a Watson computer and can instead run it on normal machinery. This way, even a middle or high school could afford such a program. His goal is to bring it down to about $15,000 to $20,000. And then it would learn. Schools would use it for a semester. It would observe the professor answering questions and it would learn those answers as it moves forward. So it becomes more and more expert every semester you use it.

Now then, faculty wonder, too: Am I out of a job? Well, probably not because there are certainly small modules that could be wanting it entirely this way, but you still need someone on the side, just like my role with my fourth-grade grandson. You need someone there to help the student when he gets hung up or where you can help encourage them. You can observe and respond to the verbal or visual signs of frustration.

So the faculty role may skew over time towards individual attention to tutor and assist students, rather than as primary delivery, drill, and assessment mechanisms for the courses, which can be done by AI.

Last year, I was at the third annual US-China AI in Education Conference at the University of North Texas and the previous tour in Beijing. One of the topics of discussion was cultural difference because programmers have an influence on what cultural values are instilled in AI programs. For example, an autonomous car is going down a small, single-lane street; there’s a telephone pole on the left and on the right is a young girl playing hopscotch. An old man steps out in front of the car suddenly. So the car has to instantly decide, Do I hit that person? Do I go left? Do I go right? Do I value the life of the child more than I value that of a senior? Do I value the lives of my occupants [inside] over those outside—because if I hit that telephone pole I may kill them? Those decisions have to be made in a tenth of a second.

The author has to help write the AI’s algorithm in a way that reflects the values, which may vary between cultures and will impact learners differently. It’s not always life and death—maybe you’re teaching calculus—but those values become very important.

AI will continue to take an increasing role in the way we each teach and learn. AI will become ever more sophisticated, so you can have more and more conditions, and programs will be able to process faster and faster.

Chelsea Toczauer

Chelsea Toczauer is a journalist with experience managing publications at several global universities and companies related to higher education, logistics, and trade. She holds two BAs in international relations and asian languages and cultures from the University of Southern California, as well as a double accredited US-Chinese MA in international studies from the Johns Hopkins University-Nanjing University joint degree program. Toczauer speaks Mandarin and Russian.