Implementing AI technology in business education

By embracing a spirit of experimentation, we can view the changes that lie ahead for our industry with optimism instead of trepidation.

By Robert Brunner

• Artificial intelligence platforms might not yet produce responses that we would mistake as human-generated, but they are learning and growing more sophisticated by the day.  
  
• To prepare students for the future of AI, faculty at Gies College of Business ask students to use AI platforms for certain assignments and then discuss what prompts their peers used to receive their results. 
• Faculty also are experimenting with the technology themselves on tasks such as grading and writing quizzes, so that they are prepared to integrate AI into their teaching.

The climate of higher education has been undergoing great upheaval over not just the past few years, but the past few months. In particular, we have seen the quick adoption of ChatGPT and the vast changes this technology has inspired in our courses.

As with many past disruptive technologies, ChatGPT has inspired great uncertainty, even fear and trepidation, among faculty. But I don’t think these are appropriate responses. On the contrary, I look upon these changes with a sense of optimism. And excitement. I believe that, as academics, we must take to heart the quote often ascribed to management guru Peter Drucker: “Innovate or die.”

With that idea in mind, I want to share my experiences and those of my colleagues here at Gies College of Business at the University of Illinois Urbana–Champaign. Our approach to adopting artificial intelligence (AI) illustrates how we can embrace new technologies and prepare students for what they will encounter in their future careers.

A Career Focused on ‘What’s Next’

My background is perhaps quite different from many people who teach business courses. It has been one of using technology, embracing innovation, and celebrating curiosity.

I am currently the associate dean for innovation, chief disruption officer, and professor of accountancy at Gies Business. But my academic training was not in business; it was in astronomy. I started my career by working on the Sloan Digital Sky Survey, a long-term project that mapped a large fraction of the night sky. During this experience, I worked primarily with big data and machine learning. Later, I accepted a position as a professor in the Department of Astronomy at the University of Illinois, where I also worked with the National Center for Supercomputing Applications.  

In 2017, I moved to Gies College of Business because I thought its programs provided a unique opportunity. Business school administrators and educators could use the university’s strengths in science, technology, engineering, and mathematics (STEM) as opportunities to really think about what is over the next hill. We could explore important questions, such as what major impacts forthcoming innovations will have on business and society. As director of the University of Illinois–Deloitte Foundation Center for Business Analytics at that time, I could contribute to how our college was educating its students and preparing its stakeholders for the future.

Originally, my role centered primarily around data analytics. But even then, administrators at Gies were thinking about bigger opportunities that were starting to take shape.

What Do We Mean by Artificial Intelligence?

Now, as the college’s chief disruption officer, I’m helping faculty make sense of the impact that other disruptive technologies, such as AI, are having on business. As we discuss the implications and capabilities of AI, it’s natural that we evaluate it through the lens of human intelligence. Consider the people we know or meet during the day. Through our interactions with them, can we gauge their levels of intelligence? Can we know, based on our perceptions of their intelligence, what it means for them to be human?

For this, people most often look to the Turing Test, which was actually called the imitation game by creator Alan Turing. Simply stated, the test asks a human being to converse with both an artificial entity such as a computer and another human, separated from each by a wall. If the human test subject cannot differentiate the responses of the AI from those of the human, we would consider that AI to be intelligent.

We are now in our own 21st-century version of the Turing Test as we interact with ChatGPT. Such platforms promise to be the most impactful technologies we have seen in recent times.

The Usefulness of ChatGPT

We can consider AI technology from two different perspectives. First, we can view its interactive interface as an amazing way to converse with some sort of artificial intelligence. In this respect, it does seem to be approaching the point of “winning” Turing’s imitation game.

Second, we can recognize that ChatGPT is a generative tool based on a large language model (LLM) that has been “trained” by scraping a huge amount of data from the internet. When we type prompts into interfaces such as ChatGPT, the AI then uses that text to start inferring its answer. When we see the answer, we can give it a thumbs up or thumbs down on how well it answered our prompt. The system will use this feedback to continue to improve the LLM over time, not just for each individual user but for the entire population of people using the same model.

In other words, ChatGPT is basically a propagating, self-training imitation game that’s just going to keep getting better at answering questions more quickly, in ways that make users happy. And this is just the beginning. We are already seeing new ways to interact with LLMs via audio and video in multimodal interactions that closely imitate how humans interact with each other.

Eventually, I believe these types of technologies will blend into the background as they become ingrained into our everyday life, in much the same way as email, text messages, web search, or social media are now no longer considered magical. The next step in the technology’s evolution could involve wearable devices that record our lives to allow an AI to help us deal with information in new and more powerful ways.

Teaching With AI

At Gies, I created an emerging technology course that I now have taught for the last four years. I tell each cohort of students that we are not teaching them to predict the future—that’s a fool’s errand. Instead, we are trying to teach them how to predict possible scenarios. If they know what the possible scenarios are, they can better prepare themselves—and their companies—for those outcomes.

The course is built so that the students don’t just learn about these emerging technologies; they actually must use them as part of their assessments. Thus, we teach the students what AI can and can’t do. We ask students to use AI such as ChatGPT to develop answers for some assignments, as part of the assignment rubric.

However, we don’t want them to just use an AI platform once and say, “Here’s my essay.” Instead, we require them to examine the process of using the technology. What prompts did they use? How did changing the prompts change the responses they received? How much do they trust those responses?

To help them answer these questions, we include peer grading in these assignments. In this case, it’s not about grading someone else’s work as much as it’s about seeing how different students approached the problem at hand. It can be extremely useful for students to know what prompts their peers used to receive their results.

Other faculty at Gies College are also bringing AI into their courses. For example, in IT for Networked Organizations, my colleague Vishal Sachdev developed a bot to serve as a support to students. It is intended to act like a member of the classroom, but with better memory. Based on the course syllabus, this bot helps students become familiar with the basic outline of the course.

Next, Sachdev developed a new bot with the same back-end infrastructure as the first. He trained this one with assignment instructions, a model submission, a rubric, and guidelines on how to respond to students. This bot can provide critiques to student submissions, feedback on possible missing elements in assignments and problems with word length, and guidance on whether students have sufficiently answered the assignment’s question.

We also have a Teaching and Learning team that supports both our in-person and online programs, examines available educational technologies, and provides faculty with guidance and leadership on how the newest edtech can be brought into the classroom. For the past couple of years, in conjunction with our Disruption Lab, our Teaching and Learning team has hosted monthly Zoom coffee hour meetings called Teaching with Innovative Technologies. During these meetings, Gies faculty share their experiences implementing specific technologies, including AI, in their courses—whether they’ve streamlined their grading with AI or have used chatbots to engage their students. These conversations are opportunities for our faculty to discuss the benefits AI tools can bring to the classroom.

Tips and Techniques

There are a few things I would suggest to others who are interested in bringing AI into their classrooms:

Know that there is a learning curve. Using an AI interface is not as simple as typing text into a Word file. At least not yet. You’ve probably heard about prompt engineering, which is crafting the best prompt to get ChatGPT to write something that is as close as possible to what you want. Many people are out there offering to teach people how to get better at this, but we have to be careful—not all of this training is relevant to higher education, or even useful.

Be willing to experiment. You likely will fail at the start, but don’t stop testing the boundaries of the technology. The beauty of using these tools is that they don’t complain if you ask them to do the same thing over and over. To get good at using AI platforms, you’re going to have to spend time with them to understand how they work and learn how to use them in your courses.

Don’t be afraid to try it in unexpected contexts. AI might be of value in areas where you thought it wouldn’t be helpful! One thing that most surprised me was how well—and how amazingly quickly—these tools could generate quiz questions based on content that I wrote for the course. This addresses one of the tasks I find to be the biggest drudgery of teaching: creating quizzes. Now, I can use ChatGPT (or another AI tool) to create 25 questions, and then I can pick the best 10 questions and edit them as needed.

With a little prompt engineering, you can customize the results. You can tell ChatGPT that you want two answers that are right, or you can ask it to generate one right answer along with an explanation tailored to the level of your students. It also will generate different answers with potential rubrics and explanations for why each answer is correct. Used this way, the tool provides a huge time savings.

Be open to new possibilities. If you think about it, it’s really amazing what generative AI platforms can do. But if you’re not open to trying them in new ways, you miss the opportunities and risk missing out.

Bigger Things on the Horizon

As academics, we must be open to even bigger things coming. Earlier, I mentioned AI video generation. While it’s not to the point that I would blindly trust AI to create new video content, we are using it to create opening videos for a MOOC I am preparing. And we are starting to look at how AI can be used to translate existing video content into new languages, all while maintaining my voice and mannerisms.

In the end, we must work with these technologies ourselves to prepare our students for the new business world to come. If our students aren’t ready for what the future holds, they could quickly become obsolete in important ways that we may not fully understand right now.

The takeaway is, if we as faculty are not staying abreast of disruptive technologies such as AI, we will face an even bigger task down the line. In short, as the famous quote says, innovate or die.

AUTHOR
Robert Brunner, Associate Dean for Innovation and Chief Disruption Officer and Professor of Accountancy, Gies College of Business, University of Illinois Urbana-Champaign