Jonan Phillip Donaldson, Ph.D., assistant professor in the Department of Curriculum and Instruction in the UAB School of Education and Human Sciences.The arrival of generative AI over the past few years is just the latest change in a decades-long evolution of higher education, says Jonan Phillip Donaldson, Ph.D., assistant professor in the Department of Curriculum and Instruction in the UAB School of Education and Human Sciences.
“Fifty years ago, higher education was mostly about knowledge,” said Donaldson, who directs the school’s fully online master’s program in instructional design and development. “Thirty years ago, it started to shift to skills.” But the jobs of the future will not require either of those, Donaldson says. “Knowledge is available everywhere, and soon enough, thanks in part to AI, there will be tools that do almost everything. My students need to develop a mindset of exploration and tinkering. There is no point in teaching them a specific tool, because that will always be changing. But if we help our students get into the habit of always trying, then they will be developing skills that allow them to flourish.”
UAB’s master’s program in instructional design and development focuses on corporate, informal and higher education learning and development positions. Donaldson’s students will graduate to jobs with titles such as “learning engineer,” “instructional design manager” or “chief learning officer.” Some of those students are teachers who do not want to leave learning but are tired of the constraints of a classroom. Many others are UAB employees who see an opportunity to expand their skills for free with UAB’s educational assistance benefit.
Emphasis on tinkering and exploration
Donaldson specializes in the learning sciences, an interdisciplinary field that was founded in 1991 with the idea that “learning is complex and messy — and it needs to be,” he said. The emphasis on tinkering and exploration in UAB’s master’s program characterizes the learning sciences approach and “is particularly relevant to the use of AI in education,” Donaldson said. “AI tools can be incredibly powerful, but their full potential is often unlocked through experimentation and discovery.” (Donaldson often refers to AI as “augmented intelligence” rather than “artificial intelligence.”)
A common assignment in many of Donaldson’s classes is to kick off a new topic, such as performance support, by “having a really in-depth conversation with AI,” he said. “The back and forth is the key. I am always telling my students that the good stuff comes after a dozen or more prompts. Then we’ll have a group discussion online and they will talk about what their AI conversations were like, what’s bothering them about the topic, what they love about it. Students tell me, ‘I get so much more out of that than a reading or a video.’”
Just "use things if they are useful"
On July 22, Donaldson will teach a hands-on workshop as part of the Center for Teaching and Learning’s “Using AI in the Classroom” series on using generative AI for systematic literature reviews.
Working as a learning engineer means constantly creating learning experiences and resources, so Donaldson’s students are doing the same. “They are making stuff a lot,” he said. “It is very generative.” And generative AI tools are very helpful for that work, he says. “I encourage them to find their favorite and get a subscription,” Donaldson said. In an interview, he demonstrates one tool that can provide narration, in his voice, as quickly as he can write a script.
“I just want my students to explore and use things if they are useful,” Donaldson said. “I have them teach each other about tools they have come across, and many of them I have never heard of, either.”
Donaldson also shares his practices with his students. Each year, he submits several papers to the major conferences in learning sciences; some are accepted, others are rejected. So “I put all these [rejected] papers, and the peer reviews I got, into a generative AI tool and said, ‘Please give me feedback on this new paper,’” Donaldson said. “It is always better than the human reviewers. Sometimes the human will just point out things that are bad, or that conflict with the reviewer’s pet theory. I show that to my students and then have them go do it to their own papers. I say, ‘Get feedback before submitting your paper and take that feedback into account.’”
Generative AI tools as "partners in the learning process"
In one of Donaldson’s classes, students finish the semester with a design or research project. Recently, a student wanted to do a systematic literature review, but could not find a partner to do the independent screening that this requires. Donaldson encouraged the student to collect paper titles and abstracts and create the criteria for what should and should not be included. Then he could give this same information to two different AI tools and get their ratings and rationales for including or excluding each paper. “I said, ‘Go to Claude, ChatGPT or Gemini, and ask it to rate each paper as a zero for ‘definitely not include,’ a one for ‘probably not include,’ a two for ‘probably include’ or a three for ‘definitely include’ and justify those reasons. Then go to another tool and do the same thing. If they agree, log that choice. If they disagree, log it as a ‘no.’ Now you have this other independent reviewer.”
Graduates of the Instructional Design and Development program “are well-prepared to leverage AI tools in their learning design situations, not as content delivery mechanisms, but as partners in the learning process,” Donaldson said. “They can design activities where learners use AI to generate ideas, test hypotheses and create new knowledge.”
Teaching with a learning science-based approach
How does training delivered by someone with a traditional instructional design background contrast with a learning sciences-informed approach? “Imagine you are tasked with delivering training on a new software platform,” Donaldson said. “Someone with a traditional background will create a structured curriculum, including a linear set of modules, each focused on a specific feature or function of the software”; deliver content-focused instruction; and assess the success of that training with quizzes and tests focused on recalling information and performing specific tasks. “This approach assumes that learning is primarily about acquiring knowledge and skills through direct instruction and practice,” Donaldson said. Someone using a learning sciences-informed approach, on the other hand, “would create a learning environment that encourages learners to tinker and experiment with the software, discovering its features and functionalities through hands-on exploration,” he said. “Learners would be encouraged to work together, share their discoveries and troubleshoot their problems collaboratively.” And the educator would assess understanding “through authentic projects or tasks that require them to apply their knowledge and skills in real-world contexts, not hypothetical scenarios based on real-world situations.”
