Instructional Designers Empower Educators, Students in Use of Artificial Intelligence Tools

This photo was created using generative AI with the following prompt: “teacher using artificial intelligence.”

AI in Higher Education

Instructional designers keep pace with the latest pedagogical and technological advances available when developing online courses. That's their job at the University of Arkansas Global Campus, but it's a tall order to stay abreast of the latest – still evolving – technology: artificial intelligence.

AI is here, and there is no going back, said Ken Muessig, instructional design manager. It is seeping into our daily lives at home, work and educational institutions. Global Campus instructional designers are working to help normalize the use of AI in the U of A classroom.

"We want to prepare students for their future careers, and in preparing them for their future careers, we need to teach them the skills that they're going to need," Muessig said. "They don't need to recall things anymore. We have the Internet and Wikipedia and other things for that, but what they're going to need to do is know how to ask the right question to get the answer that they need, and that's prompt engineering, as it's called."

Mandy Eppley agreed. She works on the non-credit side as a senior instructional designer in the Professional and Workforce Development division of Global Campus.

"We have to train students how to use AI," Eppley said. "They are going to learn how anyway, and it will be a lot better with instructors and IDs leading the way."

The Global Campus supports the academic colleges and other units in the development, design and delivery of online programs. Enrollment in online degree programs at the U of A has grown by 153% over the past 10 years, according to the Global Campus Annual Report. The instructional design team assists with the development of both credit and non-credit programs offered by the U of A. The Professional and Workforce Development division reported 2,205 total course registrations for fiscal year 2023.

A Tool to Use

In addition to learning to use various large language models based on machine learning as they are introduced to the public, the instructional designers are paying close attention to the capabilities offered by Blackboard Ultra, the learning management system used by the university.

Instructional designer Camie Wood said Blackboard Ultra recently released AI generation tools. The instructors she works with are starting to use image generators, too. Instructors can use AI to create grading rubrics, learning objectives, syllabi, and assessments, she said, as well as simply come up with ideas for all of these teaching components.

"We try to really position it as a tool instead of the teacher," she said. "It is about leveraging AI as a tool instead of doing the same things we have always done."

In her work designing non-credit courses, Eppley and other instructional designers on contract create course content in partnership with people who are called subject matter experts. They design courses in subjects such as IT readiness, drone flight crew fundamentals and advanced manufacturing.

"We use AI quite extensively to build the content initially for our contract IDs, but that makes our subject matter experts all the more valuable because they have to go through line by line and make sure everything is correct," Eppley said. "It has been an incredible tool for us because it takes away so much of the grunt work we used to have to do. I would say it has saved us weeks, if not months, in production time."

The subject matter experts are key to developing high-quality content, she said; AI programs can't be relied on without human oversight and interaction.

"You have to have a really good subject matter expert," she said. "That's the key. If you don't, then it doesn't work and you are putting out content that is not good quality and you can run into trouble. So, we have to be picky with our subject matter experts and they have to spend a lot more time. Obviously, they also give us content."

Boosting Student Learning

When instructional designer James Martin is working with a U of A instructor to create or modify a course, he describes the benefits of low-stakes quizzes to gauge learning. A former high school English teacher, Martin said these quizzes work well to help an instructor know whether students are learning the content.

"And AI is a fantastic quiz-creation tool," Martin said. "I am a huge fan of reading quizzes of the right sort."

Martin has written posts for LinkedIn about his own experimentation with various AI tools and quiz creation. One program made a rookie mistake, he said, putting the answer to the first question into subsequent questions. Another program provided an answer key without being asked.

James Martin's Experiments

Building Reading Quizzes with AI

As an instructional designer, I’m a fan of low-stakes, objective, reading quizzes. But building them is time consuming. Well, at least, it used to be. In the age of AI, it’s a snap. I decided to put ChatGPT and Claude to the task.

David Hume, Skepticism, and AI

Continuing with the #AI in #education theme, here’s another experiment in pitting OpenAI’s ChatGPT and Anthropic’s Claude against each other, this time on the subject of David Hume’s “An Enquiry Concerning Human Understanding.”

Pi Joins the Reading Quiz Generation Comparison

A colleague introduced me to Pi.ai from Inflection AI yesterday, so I decided to test its reading quiz creation chops.

Faculty are open to other uses of AI, he said.

"We used it in one class to generate objectives because that was the gap we had," Martin said. "We had lots of great lessons, but no real objectives, not stated ones anyway; there were implicit ones but not explicit ones."

He has also talked with instructors about the ethical implications of using AI in their courses and how to document its use.

"I tell instructors there are ways we can use it ethically," he said. "We can put it in our credits. We can say quizzes are generated by AI in the credits."

Instructor Oversight

As in the professional and workforce development courses Eppley described, courses in the degree programs must be carefully reviewed by instructors. Instructional designers said instructors must take a close look to be sure what AI generated makes sense and is accurate as well as furthers their course and learning objectives. AI is known to produce results with inaccurate information at times, the team said, which are called hallucinations. Like many tools, AI is only as good as the user.

"I didn't know about this hallucination thing for real until I asked it to give me a list of citations to get started on a topic I didn't know well," Martin said. "Then, I go try to find them and half of them don't exist."

In his experience, Martin said, often people are better at editing content than they are at generating a first draft. AI can be used to generate a quiz on a piece of reading material in a matter of seconds, after which instructors can use their time to verify the information is accurate and will help students learn.

"It is easy for instructors to look over what AI produced," agreed Bob Hinton, senior instructional designer. "They have to know, is that the correct answer? We know ways it can be a time-saver. I have shown a few instructors how in Blackboard it can create a discussion using Bloom's taxonomy."

Bloom's taxonomy is a framework used in education to classify educational objectives and skills into different levels of complexity. It was developed by Benjamin Bloom and a group of educational psychologists in the 1950s. This is according to a query to ChatGPT and verified independently by an article on the TIPS page on the U of A website.

Instructors and students must also understand, the instructional designers explained, that the quality of information AI produces relies heavily on the accuracy and specificity of the prompt, or question, the user gives it.

Muessig and two of the other designers took an online course through Vanderbilt University about prompt engineering.

"How you form the question radically changes the results," he said. "That course was very, very interesting and has changed a lot how I interact with AI and ask it questions."

Learning to write good prompts also helps instructional designers create content for different audiences, Eppley said. Prompts can be changed to get content that matches any level of instruction.

"Some of our classes we write for learners on an eighth-grade level and some we write for professionals with master's degrees," she said. "Sometimes, we tailor our questions and we do several reiterations of it and ask follow up questions, but that really helps us save a lot of time as far as our audience and making sure that it's written the way they need it."

What About Cheating?

Instructors are concerned about plagiarism, Wood said. Students can commit plagiarism, the act of copying someone else's work and passing it off as your own, and other types of academic dishonesty with or without AI.

"That's when we talk about the quality of test questions and what you are asking students to do," she said. "We balance that with how much time you have to grade."

Martin suggests instructors should look at ways to use AI effectively.

"I'm not saying we don't worry about academic integrity," Martin said. "We will always worry about academic integrity but to say it will be an academic integrity problem and shut it down, that has never worked."

Several academic and leadership offices at the U of A are developing guidelines on the use of artificial intelligence in academic work. The early guidance offered has emphasized informing students when they can and can't use AI, Hinton said.

"It will be assignment-specific," he said. "Maybe this assignment, you don't use AI, but further down on other assignments, these are the specific instructions on how you can use AI."

The Wally Cordes Teaching and Faculty Support Center has hosted informational events for faculty, and University Libraries offers a research guide on the topic.

Students' understanding of ethical use can be bolstered while also teaching them to be savvy consumers of content generated by artificial intelligence, Martin said. Instructors can assign lessons requiring the use of an AI tool and carefully explaining what's acceptable.

"They can tell students, pick your favorite AI, ask a question, and we can discuss the answer," Martin said. "We can put students in the driver's seat, talk about including good documentation. What exactly is your prompt, discuss and evaluate, in addition to learning how to ask good questions."

---

Postscript: After some initial research was done for this story, interviews were conducted and notes were cleaned up, this writer used ChatGPT for an organizational assist with this prompt:

Imagine you are an experienced blog writer. Take the following list of topics related to artificial intelligence and separate them into two groups with a header on each that emphasizes their similarities or connectedness. The purpose is to utilize this content in two blog stories about AI in higher education. Here are the topics: …

Some of the resulting suggestions I used; some I did not.

Online Learner Blog Home

 

Your
Thoughts?

Related Articles

• Faculty Member Describes Practical Uses, Dilemmas of Artificial Intelligence
• Instructional Designers Address Challenges, Promote Equity in Artificial Intelligence