‘Question of our time’: Colleges grapple with AI’s impact on student learning

“Why are you here?” Jessamyn Doan Ewing asked a room of roughly 30 first-year students at Agnes Scott College.

Doan Ewing, director of academic advising at the all-women’s Decatur college that was founded in the late 19th century, was asking profound questions provoked by a disruptive 21st century technology. What, for instance, is the point of college?

The discussion that followed was part of a class on artificial intelligence, added this fall to the private school’s curriculum. Two weeks of the seven-week course, mandatory for all incoming students, are now focused on AI, including how it works, the ethics of its use, and ground rules for when it’s permitted in their education. Other schools across the state, such as Georgia Tech, Georgia Southern University and Georgia Gwinnett College recently added mandatory AI elements to their curriculums; GSU requires all incoming students take a 90-minute “AI fundamentals” course online, while required computing classes at Tech and GGC now include AI components.

The classes at Agnes Scott address a broader challenge encountered by universities nationwide: reminding students that learning itself is a difficult but worthwhile endeavor, one that is lost if they rely too heavily on AI.

“Your brain is going to learn best when it has to work hard to figure out the answer, and AI can sometimes short-circuit that,” Doan Ewing told The Atlanta Journal-Constitution minutes before the class began. “[Students] want to feel equipped to tackle the challenges that our world throws at them, and that requires learning.”

Virtually all universities across Georgia and the country are grappling with what AI means for students, professors and the entire higher education enterprise. Recent Lumina Foundation-Gallup polling shows roughly 60% of college students are using AI on assignments weekly, while a different survey found 90% of professors are concerned AI use will diminish their students’ critical thinking skills. Its prevalence, and its ability to quickly complete assignments, raises existential concerns for colleges everywhere.

If students can bypass the learning process by offloading their coursework to a machine, what is the purpose of a university?

“In many ways, this is the question of our time,” said Lauren Klein, an Emory University professor who leads an AI research initiative for the school. “The introduction of AI into students’ lives has really almost rack-focused the purpose of the college classroom for me.”

It is no longer enough for colleges to simply teach students the course material, Klein argues. In the age of AI, faculty must ensure their students learn how to learn.

Depriving students

This summer, a committee at the Massachusetts Institute of Technology, a place that has produced some of the smartest minds in the tech world, issued a call to action. In a 40-page report examining the role of AI in learning and teaching, the committee determined what should worry the school most about AI is that it can “deprive students of the opportunity to learn.”

It offered recommendations on how the university could adapt, saying that with “intense societal focus on credentials and grades,” students who want to maximize their grade-point average are tempted to use AI without actually learning the material.

Klein said AI has made it possible to get a college credential “without having that experience of learning and of expanding your mind.” Similarly, the MIT report determined educators must help students understand the most important product of their education is “not a GPA or a diploma but themselves,” adding that students must not only be taught what to learn, “but also how they should learn and why learning matters.”

When the grade isn’t perfect

A few months removed from graduating high school, two dozen Agnes Scott students were being taught the science of learning.

After comparing the neuroscience behind different study strategies — such as flash cards, rereading notes, or studying with music — Doan Ewing brought AI into the conversation: how it works, what it’s good at, what it’s not, and how it could help (or hurt) their learning experience.

Students shared stories of their previous AI exposure. In one high school class, a student said, classmates would put assignments into a chatbot, then copy and paste the answers.

Why is that a problem?, Doan Ewing asked.

“You’re not using your brain anymore,” a student replied. “Once you go into the real world, you don’t have any skills to do anything.”

The conversation led Doan Ewing to ask what they want out of their time in college. Is their objective merely to get good grades and a degree?

“A class where you got a B or a C, but you really wrestled with that material and you made a lot of progress, can still be a really successful class even when the grade isn’t perfect,” she said, adding that employers want students who can problem solve, think critically and construct arguments.

“That’s what you’re learning to do in the classroom,” she said. “The things that feel really, really hard are actually the things that are more effective for helping you learn.”

No cookbook

Some schools are avoiding AI altogether because “it does impact your ability to have your critical thinking skills,” said Spelman College President Ayanna Howard. Other schools want to make every student fully AI competent.

“The [schools] in the middle, where we’re sitting, we’re struggling with this,” said Howard, who began her role as president in August. So far, the private all-women’s historically Black college near downtown Atlanta has primarily left it up to individual departments to determine how to approach AI, although it recently brought on a special adviser to the president, tasked with examining the school’s AI strategy.

“It requires professors to be creative as well because this is a new tool, this is a new technique, and there is no Cookbook 101 on how to do this,” said Howard, the author of a book released this month that explores the challenges and opportunities of AI.

“There is a little bit of having to build a plane while you’re flying it,” said Delana Schartner, assistant provost for academic policy at Georgia Southern University. School leadership has directed every academic program to incorporate AI into their curriculum, the goal being to eventually have all Georgia Southern graduates be AI literate, giving them an advantage in the job market, according to provost Avinandan Mukherjee.

“I hear a lot of students saying, ‘Oh, I want to use [AI] for everything,’” said Schartner. “We really want to make sure that they use it responsibly, and they know how to use it correctly, and they understand the outputs that they’re getting.”

Mukherjee said AI is being used at the school as “an enabler to to critical thinking, to analytics, to learning, to everything we do in the academy.”

Without that understanding, humans can fall prey to what researchers have dubbed “cognitive surrender,” a phenomenon in which users stop constructing their own answers to questions and blindly accept the judgments of a chatbot.

“The great risk,” said Joel G. Thomas, an assistant professor of psychology at Agnes Scott, “is falling into the delusion that AI can tell you what and how to think.”

Nothing in their heads

Sophia Kaye, a third-year business major at Georgia Tech, said that while AI can be beneficial, its use can come at a cost.

“A big con is that it just takes away from critical thinking and problem solving, and being able to actually think through things alone,” Kaye said.

It’s also made it easier for students to cheat, while simultaneously blurring the line of what constitutes cheating.

“If you do use it you’re wondering, is this breaking a rule? Am I doing something wrong?” said Christian Cain, a grad student at Georgia Tech.

An academic honesty report from the University of Georgia shows during the 2024-25 school year, the number of cheating accusations nearly doubled. And for the first time, the annual report included an AI datapoint: 51% of the incidents were “alleged to be related to the use of artificial intelligence.”

Related Story

---

College students are cheating. Professors are struggling to stop them.

March 18, 2026

Mahalia Saavedra, a first-year student at Agnes Scott, said many of her schoolmates are averse to AI. While she sometimes uses it to study — by explaining her understanding of a concept to AI, the chatbot can identify parts of the concept she needs to study more — Saavedra is concerned about long-term implications if students outsource their learning.

“I worry that the world and the people in it are going to have significantly less substance,” she said. “Nothing in their heads. But they know how to use AI. They know how to do their job or whatever by prompting something else to do the work for them.”

The college experience

What is a university for?

The grades, the degrees, the credential toward a well-paid career are all part of the college package. But experts also refer to the value of an educational “experience.”

It’s a school’s responsibility to help students understand they’re in college for “the experience of thinking and of thinking in community, of getting things wrong in addition to getting them right,” said Rachel Bowser, the vice president of academic affairs at Agnes Scott.

“There should be some C’s here and there, probably, and some deadlines missed, and some stressful things that happen,” she said. “I do think it’s incumbent on us to remind students that it’s not supposed to be fast and that it’s not supposed to be easy.”

If colleges don’t provide the resources to ensure learning happens, said Klein, students may still get the degrees, but they’ll miss out on something larger: the rare experience of “what it is like to expand your mind, to realize that there are always more things to learn, and that by learning those things, it makes you feel more alive.”

Going old school

Amy Bruckman has experienced many changes in higher education since she first started teaching at Georgia Tech in 1997. But AI, according to the computer science professor, has created a moment of “unprecedented disruption.”

“I have never seen anything like this before. We are all looking around, questioning the basics of what we have been doing for generations,” said Bruckman.

Like so many professors, she wants to be sure learning is still happening in her classroom. That’s meant Bruckman and colleagues have had to get creative.

She and others have replaced online exams with Bluebooks, requiring students take tests in class with pen and paper. Many have eliminated “reading reflections,” finding it was far too easy for students to simply have a chatbot write a summary of assigned passages. Others are turning to oral exams.

With some professors, “you’ll turn in an assignment, and then they’ll call you into their office and then briefly ask you questions about it,” said Cain, “just to make sure you understand what you’re talking about and it wasn’t purely from ChatGPT.”

In her English class at Emory, Klein banned screens and bought notebooks for all her students. The strategy not only removes the possibility of AI use, but Klein cited research showing note-taking by hand leads to better retention. In another Klein class, students are learning to train a language learning model, an assignment that can’t be completed without AI. But they’re required to present their step-by-step process to the class, explaining how they worked with the technology to complete the assignment.

Georgia Tech recently released an “Academic AI Playbook” to give faculty ideas on how to adapt to AI use. The collection of 67 stories from GT professors offers firsthand accounts of what’s worked and what hasn’t.

The 280-page document concludes with 20 lessons. One was to make AI policies explicit and reasoned: “Blanket bans and blanket permissions both failed. What worked was assignment-level specificity paired with an articulated rationale.” Another emphasized human connection, citing a class where instructors converted an individual assignment into a group project. They found by emphasizing human-to-human interaction, the shift “improved the overall quality of submissions while effectively precluding AI-generated work.”