Let’s Give AI a Hand: Student Collaboration with AI Through Foundational Expertise
Dr. Craig Depken, II
Human/AI Collaboration, Foundational Knowledge, Critical Thinking, Subject Matter Expertise, Balanced Approaches, Student Success, AI Literacy
Dr. Craig Depken, a Professor at the Belk College of Business and Chair of the Economics Department, never imagined that a simple conversation with AI would lead him on a journey of both wonder and caution. His first encounter with the power—and pitfalls—of artificial intelligence began one crisp autumn afternoon when one of his master’s students knocked on his office door with an unusual research proposal.
Intrigued yet skeptical, Dr. Depken agreed to explore the idea together. With nothing more than a curiosity to drive them, they turned to a popular AI tool for insights on the emerging topic. To their astonishment, the AI confidently returned a list of authoritative-sounding papers, complete with detailed journal names, titles, and even co-authors, some of whom hailed from far-off European institutions. At first glance, it all appeared credible. However, as Dr. Depken soon discovered, the papers were nothing more than elaborate fabrications—a vivid example of what the tech world has come to call “hallucination.”
How did an AI that could mimic expert language so convincingly fail so spectacularly on factual details?
That moment marked a turning point for Dr. Depken. As he ascended from his second-floor office to a bustling third-floor classroom later that day, he couldn’t shake the uneasy feeling that technology, while remarkably capable, was far from infallible. His mind raced with questions: How did an AI that could mimic expert language so convincingly fail so spectacularly on factual details? Determined to uncover the mechanics behind the phenomenon, Dr. Depken embarked on a deep dive into the realm of large language models.
In his journey of discovery, he began to understand that AI operates on probabilities and patterns—sifting through an immense corpus of human knowledge to generate responses that seem logical on the surface. Yet, without a robust foundation in the subject matter, even the most sophisticated model could lead one astray. “It’s a lot like using a desktop calculator,” he mused, “if you don’t know the math, even the best calculator won’t help you solve the problem.”
Dr. Depken’s exploration didn’t stop at theory. He began integrating AI tools into his academic work, treating them as a kind of research assistant. For instance, he experimented with advanced models like Claude to help troubleshoot persistent coding challenges in his research projects. On one occasion, after months of grappling with a stubborn blind spot in his econometric analysis, he fed his code into the AI. To his amazement, the tool not only diagnosed the problem but also proposed an innovative alternative approach—one that he hadn’t considered before. This unexpected breakthrough reinforced his belief that when wielded by a knowledgeable hand, AI can be a powerful ally.
However, Dr. Depken’s enthusiasm was tempered by the realization that the very features making AI impressive also posed significant risks. In the classroom, he observed how easily students could fall into the trap of over-reliance on AI-generated content. One professor recounted assignments where demand curves inexplicably sloped upward—an error that indicated a superficial engagement with the subject matter. For Dr. Depken, these experiences underscored an important pedagogical challenge: ensuring that students develop a strong foundational understanding before leaning too heavily on AI.
As he continued to navigate this evolving landscape, Dr. Depken began advocating for a balanced approach in both education and research. He envisioned AI not as a replacement for human intellect, but as an indispensable tool that, when used judiciously, could amplify productivity and spark innovation. At the same time, he warned his colleagues and students about the dangers of letting AI “feed on itself”—a scenario in which unchecked reliance on machine-generated information could lead to a cascade of inaccuracies.
Looking ahead, Dr. Depken remains both excited and cautious about the future of AI. He sees it evolving into an even more capable research assistant, one that demands from its users not just technical proficiency, but also critical thinking and deep subject-matter expertise. In his view, the true promise of AI lies not in replacing human ingenuity, but in complementing it—provided that educators, researchers, and students learn to ask the right questions and interpret the answers with a discerning eye.