Opening up about OpenAI with Professor Nicola Bianchi

By: Sammy Tartell, Staff Writer

As generative AI sweeps the nation, almost every syllabus reflects the same shared sentiment: “any use of generative AI will be viewed as a potential honor code violation.” But as AI becomes more sophisticated and ubiquitous, universities have begun to view it more holistically, reconsidering it as a tool for learning versus a mere substitute. First year MBA students saw this in action this fall; for the first time in Kellogg’s history, a core class was being built around AI. With the pilot now well underway, I sat down with DECS 431 Professor Nicola Bianchi to find out how it’s going.

Professor Bianchi is a labor economist and Associate Professor in the Strategy Department, teaching Business Analytics (i.e. "DECS") II since joining Kellogg in 2015. In the professor’s 10 years of teaching this class, this is the very first time the course has seen material change in software. Like all core classes, many electives depend on DECS 431, so transitioning from classical software to AI software was not a decision made overnight. It started with a  high-level discussion about the best way to teach data analysis at the fundamental level. Old-school Stata was great for its user-friendly interface and a big step up from its predecessor, Microsoft Excel, but generative AI was disrupting the business world, so controlled exposure in the classroom felt like the responsible way forward, Professor Bianchi shared with The Kelloggian.

Professor Bianchi teaching RVF plots using ChatGPT in DECS 431, Section 31

Still, he’s upfront about his early reservations around bringing AI into the classroom. He and the team of professors teaching DECS 431 dedicated much of their summer to testing the new software against the coursework to ensure a smooth rollout in the classroom. And then, about a week before fall quarter, OpenAI introduced ChatGPT 5, replacing the O3 model Bianchi had gotten to know so well. Despite the many hours he’d invested in bringing this technology to his students, he was ready to reverse his decision at any given point if he felt uncomfortable with the implementation. Fortunately, the professor’s mid-quarter evaluations were as positive as, if not slightly better than, last year’s, indicating that the introduction of AI had not led to any glaring issues from the students’ side.

Having worked so closely with AI for so long, the professor had a few key takeaways to

share. His biggest one: only use ChatGPT for things you already understand… use it as a

powerful assistant. “There is a lot it gets wrong. The output can often be poor quality despite

seeming intelligent,” he explained. “But you can only catch these mistakes on material you

already know.” 

Prof Bianchi described ChatGPT as a glorified calculator: it’s great for the basics

but becomes a problem when applied to more advanced topics. To avoid problems with the

latter, our prompts need to be hyperspecific. To be able to specify, you need

to already know what’s important.

He closes out the interview by reflecting on what it means to be a student today: if he were in

our shoes, the truly scary thing would be to rely on AI for learning new material. That, he says, is “a very bad position to be in,” and one he would never accept in his own work or life more broadly.

As Kellogg experiments with integrating generative AI into its core curriculum, one thing is clear: the technology may be new, but the responsibility it demands is timeless. AI can sharpen skills, accelerate analysis, and open new doors… but only if students stay grounded in their own understanding. As Professor Bianchi reminded me, the real work of learning can never be outsourced. And in a world increasingly shaped by algorithms, that grounding might be the most important lesson of all. Read More Faculty Features by Sammy Tartell: An Interview with Professor Carter Cast Coffee Chat Crumbs with Loran Nordgren