The AI Paradox: What it really means to use AI in university
As AI becomes a constant study companion, students and professors grapple with convenience, learning and what might be lost when machines begin to think alongside us.
A sticky note that reads "The AI Paradox" among a pen and a textbook. Photo credit: Samara Murray
Sitting on his bed at 6 p.m. on March 20, a MacEwan University Bachelor of Science student was running out of time.
He had just come home from Calgary, and the STATS 151 assignment due in less than six hours was not yet completed. The computer on his lap showed tabs stacked across the screen — YouTube, Google and Spotify — and the indie music playing through his AirPods blended into the background.
He struggles through question one. He makes it through question two. The deadline pushes down on the student, who wishes to remain anonymous to avoid an Academic Integrity violation. By question three, he opens one of the many ChatGPT+ tabs on his computer. His cousin pays the plus subscription, and the students use it for the unlimited photo uploads. With the assignment document uploaded, the student prompts the LLM for the formulas to use.
He only uses ChatGPT as a second teacher. It is much easier to ask a machine for help than to wait for a professor’s email or talk with a friend.
He never directly copies the generated answers, and he “never supports anyone” copying answers, he says later.
When AI is helping him learn, he feels productive. He wants his degree for the knowledge and skills. He tries to avoid using AI, and he shakes as he recalls this behaviour. Nervous about the interview or admission, eye contact is secondary, as he reveals he uses AI at a 9/10 level.
The assignment is submitted at 11:50 pm.
Calin Anton, a computer science professor at MacEwan University, says his department aims to help students prepare for their first jobs. The number one priority is teaching students how to learn for themselves. If they were to sit on their degree skills in the workplace, “they’ll be dinosaurs in five years,” he cautions.
Anton uses AI’s fast-paced growth and efficiency to help students keep pace with the market and learn new technologies as they emerge.
The evolution is so quick that he will tell his students, “Whatever I’m telling you today may be irrelevant next week.” He encourages his students to use LLMs to stay up to date rather than asking him for help.
And they do.
Anton sways in his office chair to explain why: “It’s more efficient.”
AI learns through algorithms that guide these machines to “think” independently and make decisions based on massive datasets. The work of AI systems like ChatGPT is much more efficient at taking in data and then spitting it out in unique forms, like essays, emails, or practice questions, than people are. It took the Bachelor of Science student only minutes to upload his assignment and receive prompted help from the bot.
Back in professor Anton’s office, a question remains. Why do we need to be as efficient as possible?
“Competition,” he says plainly. In computer science, work is done in teams, and in a team, “you need to be at least as efficient as everybody else.”
For the rest of the market, the efficiency of AI systems follows the money. And the money is in acting like a teacher to students. It’s in the video of Will Smith eating spaghetti from last year that made university student Sienna Wilchak fear what people could do with AI technology.
“These market forces mediate every aspect of our life,” says Alain Beauclair, a philosophy professor at MacEwan University. Outlook’s AI writes automated responses within his inbox, and he jokes that he won’t have to write emails eventually. That one day it will become automated replies back and forth. “…They define how we live day to day.”
The global AI market is expected to reach approximately 4.2 trillion dollars by 2030. As these advanced tools become pervasive in the world’s technological landscape, more people cannot function without them.
Jacqueline Baker, an English professor at MacEwan and a Canadian fiction writer, says students are often tempted to use AI because of time pressure and convenience. “It’s just right there, a click away… It’s too easy.”
Learning and the Role of University
A man wielding a blade enters a Mac convenience store. It is January.
“Give me all this money,” he tells the clerk. “And give me a pack of cigarettes.”
The General Occurrence Report facts stare at her from the computer screen: Witnesses and CCTV footage indicate that the suspect fled the scene.
A makeup bag sits on the cluttered black-wood desk, along with the Criminal Code of Canada and the Percy Jackson novel she’s been itching to read since Christmas. It’s 11 p.m., Monday, March 16. While her parents sleep and her brother yells at video games through the thin walls, 18-year-old Sienna Wilchak needs to work; the “two hours left” countdown on her assignment due date is ticking.
As a first-year student in the Police Investigative Studies Diploma at MacEwan University, Wilchak picked a program that would guarantee a job and also provide societal change. “When you think about the police, you don’t think about really great people,’ she says. “I wanted to try to change that in the future.”
Parsing through the assignment rubric, the criminal code, lecture notes and samples, “I just want it to be over,” she thinks. But Wilchak needs to make sure it is completely right before submitting, tediously finishing up the small details. She has a specific way of reporting on dates and times, so specific that the help of artificial intelligence never crosses her mind. “AI gets so much wrong about anything,” she says later.
After a few hours, she submits just before 11:30 pm. The salt lamp on her desk glows.
To Wilchak, the concept of outsourcing work at the expense of learning and cognitive growth is crazy; “There’s no other word for it.”
Professor Beauclair’s research is on moral imagination, that is, the role our mind plays in leading a flourishing life. To him, genuine learning is the internal process of adopting a perspective different from how one has previously viewed the world. It’s a cycle of removing your own assumptions, understanding them, and checking them.
AI doesn’t encourage any of that. “It’s the most sycophantic system in the history of humankind,” he says.
According to a 2024 survey by the Digital Education Council, 86 per cent of students across the globe reported regularly using generative AI in their studies. Yet despite the widespread use of AI, more than half of students said they do not feel adequately prepared to use the technology effectively.
“Recording things in your memory isn’t learning,” Beauclair says.
The institution of universities and colleges is fundamental to the learning process Beauclair describes. If they were merely the means of providing students the materials to accomplish tasks, AI “could replace universities in a second,” says Beauclair.
“But the purpose of education is to become human,” he says.
What We Lose
MacEwan’s only library on campus is quiet if you know where to be. On the third floor, students are on laptops, writing in notebooks, or scrolling, and the light is cool from the early sun. There is learning, of course, but the soft chatter of connection feels as much a part of the space as the bookshelves.
A building away, Beauclair describes what colleges try to direct students towards: “to have a common fund of meaning that we can explore, enrich, maybe transform—as a community.”
“When you outsource that to some kind of AI machine,” he says, “you’re divesting yourself of work that’s proper to your own humanity.”
If you were to sit near the large window in the library that spans the third floor to the second, the noise from below flows like a river every hour at minute: 20 and:50 when classes end. Friend groups talking animatedly, a man’s voice saying “ChatGPT”; words that register only to be gone again.
Professor Beauclair continues, “What dances around and sells itself in the name of convenience is actually a life that is horribly and wholly structured.”
What do we lose then, as people, if we live for convenience?
In his office, Beauclair responds before the question is fully out. “Everything.” He is breathless and resigned when he continues. “You lose what it is to be human.”
—
When computer science professor and researcher Anton was a young boy, his father taught him how to mentally add three- and four-digit numbers very quickly. He now senses that he must have had some sort of pattern recognition back then that allowed him to do this.
Then he went to school. “I learned how to do it…” he puts up finger quotes, “properly.” Learning this way made him less efficient at addition; he lost the innate ability his father had shown him.
The pressure to adopt a single global model of understanding through LLMs will force humanity to be the same, “like a robot,” says Anton.
When will that happen?
Anton doesn’t have an answer. “I hope I’ll be dead by then.”
—
In the summer of 2025, Sienna Wilchak was working at Canadian Tire when a woman came up to her in the Gardening section of the store. The woman was wearing a trench coat and glasses, likely in her mid 40s, and had her phone out as she approached.
Excuse me. “What kind of dirt would be best for this certain sort of plants?”
Sienna didn’t know the answer, but tried to help anyway. “Oh, we have a lot of different sections, let’s go talk to a seasonal associate.”
The woman stopped her before she could go get someone to help. “No, it’s okay. I’ll ask ChatGPT,” she said, and then walked away.
The Bachelor of Science student believes that the future of AI use will produce “A lot of dumb students,” and isn’t proud to be part of the system that will take society there. But he feels he has to.
“If I don’t use it,” he says, “There’s a high chance I will be left out of the population who is using it.”
To say that the choice of using AI or not is personal removes the fact that “our lives are mediated by these tools that we never chose or opted in on in the first place,” says Beauclair.
“It’s not saying that we’re completely helpless,” he amends. “But we’re the next worst thing.”
—
The lecture hall in the heart of the Robbins Health Learning Centre at MacEwan is dead. There are 40 students enrolled, but extremely few attend class. The Bachelor of Science student is one of them.
It’s Intro to Physics. The professor scribbles calculations on the whiteboard: it’s the angular acceleration unit. The student tries to understand. The calculations are erased. New ones take their place. He doesn’t catch up to the professor’s explanation.
Now he’s lost, and he doesn’t understand the origins of the calculations that he is supposed to be learning. The professor moves on.
Sneakily, so the professor can’t see his phone, the student takes a picture of the board and prompts ChatGPT to explain.
“Hey, where do you think this number is coming from?”
The student feels pity for the professor; he’s just trying to do his job. But the student is just trying to learn.
