Why Darden Is Moving AI Into the Core MBA Experience

Earlier this year, a majority of the First Year MBA students in a required strategy course at the University of Virginia Darden School of Business missed the same quiz question.

Most selected the same wrong answer.

At first, Professor Jared Harris thought the issue might be the question itself. Then a teaching team colleague entered it into ChatGPT.

The chatbot confidently produced the same incorrect response many students had chosen.

The next day in class, Harris projected the AI-generated answer onto the screen.

“I basically said, ‘This is the problem,’” Harris recalled later. “How do you know if GenAI is feeding you something wrong, or if it’s doing something efficient and valuable for you?”

The room fell silent.

” You could have heard a pin drop in there.”

The point was not that students had cheated. This year in the course Harris designs and oversees, students were explicitly allowed — even encouraged — to use generative AI for preparation, research and assignments. The problem was that many accepted the AI’s answer without interrogating it.

Similar scenes are now playing out across business schools.

AI is rapidly becoming part of the MBA curriculum. What remains unsettled is whether schools are teaching technical fluency, human judgment or some uneasy combination of both.

A recent Poets & Quants analysis of 20 leading MBA programs found Darden among the schools with the largest number of AI-related courses, second only to Stanford. But the more revealing finding was that most AI instruction across top business schools still lives in electives. Of the 166 courses identified, only nine were part of the required core curriculum.

At many schools, students still opt into AI education through electives. At Darden, however, faculty have increasingly pushed it into the required First Year experience.

Professor Harris leads a strategy class at Darden. (Photo by Caroline Mackey.)

‘You Cannot Run AI Out of the Classroom’

Several years ago, shortly after ChatGPT’s initial release, Harris and Professor Mike Lenox — now serving as interim dean — independently tested the early chatbot against Darden’s existing First Year strategy exam.

“With about 20 minutes of playing with it, ChatGPT produced an answer that would have clearly passed,” Harris said.

At first, Harris thought the challenge was relatively narrow.

“I thought the task was just a simple redesign of the exam so students couldn’t phone it in and use AI to pass,” he said.

Eventually, he came to a different conclusion.

“You cannot run AI out of the classroom.”

That realization has spread unevenly across higher education. Some professors still prohibit generative AI entirely. Others allow limited use with disclosure requirements. A smaller group has begun integrating AI directly into teaching and assessment.

At Darden, Harris and his colleagues pushed further.

“We leaned into AI 100% in core strategy,” he said. “That made us an outlier.”

The approach was not universally embraced.

Darden has confronted similar questions before.

In a recent conversation reflecting on the future of business education, Dean Emeritus Bob Bruner — a widely respected teacher in the Socratic case method who joined the faculty in 1982 — recalled arriving from Harvard with an Apple II computer.

“This caused great consternation among my senior colleagues,” Bruner recalled, drawing laughter from the audience.

He said the memory now echoes in debates over artificial intelligence and other emerging technologies.

“We’d better figure out how to use AI,” Bruner said of Darden, “and I think we will, because we figured out how to use the Apple II computer.”

In a recent blog post on academic integrity and AI, Lenox argued that many institutions are still approaching the technology primarily through restriction and enforcement — returning to handwritten exams, proctored testing and software tools designed to detect AI use.

But Lenox suggested those efforts alone will not solve the problem.

“We are fooling ourselves if we think that the solution to AI-enabled cheating is stricter rules and enforcement,” he wrote. “The technology will always outstrip our ability to restrain.”

Teaching Judgment, Not Just Tool Use

For Harris, the challenge is balancing “two concurrent objectives.”

“One objective is still timeless,” he said. “I want students to emerge having genuinely learned something — principles, judgment, expertise.”

And that means more than simply learning how to write AI prompts.

“But the other imperative for any relevant and elite business school in today’s world is to prepare students to actually work and thrive in today’s workplace, already saturated with AI tools,” Harris continued. “And that includes an increased understanding of how to interact with AI.”

For Harris, the real divide is not between students who use AI and those who don’t. It is between students who accept AI answers at face value and those with enough expertise and judgment to challenge them.

That, he argues, is why AI exposure cannot remain confined to electives.

“These learning moments shouldn’t just happen to students who choose to take an AI elective,” Harris said. “Everyone should have that moment.”

The moment he means is the silence that followed in his classroom after students realized that ChatGPT had confidently argued for the same wrong answer many of them had accepted uncritically.

The technology sounded authoritative. The responsibility for recognizing the mistake still belonged to them.