*Editor’s Note: The “Views from NAU” blog series highlights the thoughts of different people affiliated with NAU, including faculty members sharing opinions or research in their areas of expertise. The views expressed reflect the authors’ own personal perspectives.
By Michelle Miller
Professor in the Department of Psychological Sciences
Dr. Miller studies memory and attention as well as effective online and face-to-face learning activities for college students. As an expert in redesigning college coursework, she has offered workshops and presented keynote addresses at a variety of universities and national conferences.
I’ll admit, when I heard the first rumblings of news on AI in late 2022, I didn’t think much of it. As a professor of psychological sciences who writes and does research on technology, I’ve seen many such trends come and go. I was equally skeptical of claims that AI would require colleges and universities to fundamentally change how they go about teaching. In my work to disseminate learning sciences to faculty, I’ve come to realize just how slowly higher education tends to adapt to any change, technological or otherwise.
But as the innovations continued to pile up, and as I started mastering the tools myself, I realized that this could be one trend that warrants the hype. Colleges and universities are starting to sit up and take notice, and scholars in the field are as well.
This has all touched off a fast-moving debate, ranging from comments posted on social media to research articles in scholarly journals. Many of the former offer scorching critique of how AI will surely cause an epidemic of cheating, eroding learning along the way. Many of the latter effuse about the positive potential for supercharging learning. And across the board, others talk about the AI skills students are going to need to thrive in their future jobs.
It is a confusing and overwhelming time, to be sure. For that reason, the first step is to try to separate out what I think are three related but distinct issues going on in AI’s impacts on higher education.
First is the issue of academic dishonesty, which tends to be the top concern expressed by faculty new to AI. College instructors have long grappled with online services such as Chegg or CourseHero, whose databases of class materials can be used to cheat. But AI takes this to a new level, given that it actually processes content in a meaningful way rather than simply searching through existing material. Make no mistake: AI can do at least a B+ job on most exams and homework questions, along with basic writing assignments.
Right now, there are few workable ways to “AI-proof” online exams and assignments. AI “detectors” are unreliable, at least in their current form, and online proctoring services (which surveil students by video or other technological means) are too cumbersome and, I believe, too intrusive to use for routine assignments and tests. I have settled on a strategy of making cheating less tempting, if not impossible, using a combination of media that is hard for AI to review and increased personalization of assignments. I’ve also begun phasing out traditional timed online exams, replacing these with more open-ended and collaborative work.
The second issue is more uplifting, although similarly challenging: How do we tap into AI’s potential for helping students learn? One major strategy I see is using it to generate quiz questions based on what students are learning. Researchers have long known that answering questions is itself a powerful enhancer of learning, helping to solidify new knowledge and even helping students develop more advanced thinking skills in a subject area. One limiting factor to quizzing for learning has been the time and effort needed to craft these practice questions. But AI is also very good at generating questions based on content you give it and can also give feedback on student answers.
Third is the question of how to best prepare students for a future in which they will live, work and learn alongside AI. This may be the hardest one to address, because so much is still up in the air as far as how professional fields will incorporate the tools.
For example, AI could be used to generate some of the more routine and technical sections of the scholarly articles that are the lifeblood of scientific research. It might be good for students to practice having AI generate first drafts of these sections, in which writing style is less important than arranging data accurately and in the proper format. But many scholarly journals haven’t yet settled on a policy for whether any proportion of AI-generated content is acceptable, leaving students and faculty in limbo.
But as more becomes clear about the future for which we are preparing students, the better placed we will be to help them move forward. In the meantime, faculty and their leadership should keep working to stay ahead of the rapidly developing AI scene, one that brings with it major challenges as well as intriguing possibilities for promoting learning.