In a panel discussion, institutional leaders and a current student cautioned that scaling support with AI shouldn’t mean limiting student choice.
“As our systems become more intelligent, more predictive, and in some cases more prescriptive, are we expanding student agency or quietly redefining it?” said John Falchi, senior vice president for higher-education partnerships at ed-tech company LearningMate. “When a system tells a student what they are most likely to succeed in, is that guidance, or is it a constraint?”
Rohan Agrawal, a third-year economics and computer science student at Arizona State University (ASU), said students will use AI for tasks associated with counselors regardless of the school’s direct involvement. For example, when he looks for information on majors and course substitutions on ASU’s website, it involves clicking through several links and submenus. He said AI can do that digging for him, fact-checking suggested course schedules and potentially finding pathways advisers aren’t aware of. For a student like Agrawal, who is interested in economics, computer science and philosophy, AI can help personalize counseling advice and find overlap in an uncommon intersection of interests.
“It’s actually one of the places where I do feel supported the most, because AI has the opportunity to ingest all the different web pages that exist,” he said.
Maja Zelihic, assistant provost and vice president of Capella University, said this kind of augmentation from the institution’s side — using software to save time digging through institutional resources — can free up an adviser’s time for relationship building.
At a smaller, online school like Charter Oak State College (COSC), Provost David Ferreira said students are not usually looking for unique combinations of courses.
“For our folks, time is money,” he said. “Give me just the stuff I know is tied to my career.”
To assist their students, COSC used AI in early stages of the admissions process to evaluate transcripts and show applicants how many credits would apply to their desired certificate or degree before enrolling. Ferreira said this use of AI is better suited to COSC’s student population, who are 36 years old on average and typically already in the workforce. Ferreira said it is important to create goals and metrics that reflect the students’ desired education experience, including career outcomes, time and cost.
Panelists said they’ve also used AI to humanize data on student engagement. At West Coast University, Provost Kirstie DeBiase said her team of student-support staff felt overwhelmed by large institutional data sets and struggled to find actionable insights. To show them how AI might help with this, she used it to create a narrative around one student in the data set, with scripts and a video that painted a picture of someone who wasn’t an outlier in any one category of concern, but whom the data collectively suggested was disengaged. DeBiase said the exercise revealed that many students fall into a middle category — not actively failing, but still relatively disengaged, having slipped through the cracks of student success services.
“Before having AI as a tool, it was just a little bit overwhelming,” she said. “We could really only focus on who’s doing really well and who’s at the bottom of the spectrum.”
DeBiase said this is key for advising staff, because students who are not failing but disengaged are likely to drop out without drawing attention.
However, panelists envisioned a downside to advisers using AI if the decisions that come from it make students feel boxed in. For example, steering students toward majors they are statistically more likely to complete or narrowing academic options may help meet institutional goals for graduation rates.
“I think it becomes incredibly important that we don’t lose sight of student agency,” said Stuart Rice, executive director of learning experience at ASU. “The data that we have around students are really ways of signaling to the broad spectrum of people who are interacting with that student where the student might be stumbling, but not to see that as a way to tighten the constraints on the student.”
