Three questions to guide thoughtful, human-centered AI use in the classroom

Instructors at Penn continue to experiment with how AI can support learning, leveraging the technology while centering critical thinking and human insights as the backbone of a meaningful learning experience. 

In the spirit of sharing how faculty at Penn are approaching this new technology inside—and out—of the classroom, the Center for Excellence in Teaching, Learning, and Innovation (CETLI) will host a panel, “AI-informed Activities and Assignments,” on April 21 as part of AI Month at Penn. The panel will feature six faculty from the School of Dental Medicine, the School of Arts & Sciences, the Penn Carey School of Law, the School of Engineering and Applied Science, and the School of Nursing showcasing how they are integrating AI into their teaching. 

Ahead of the event, Catherine Turner, director of faculty programming and pedagogy at CETLI, offers three fundamental questions to keep in mind when weighing incorporating student AI use in a course. 

1. What are the learning goals for your students? 

Ask first: What do you want students to get from your classes? 

Knowing what students should be learning and the process to accomplish that will allow faculty to better design their assignments, says Turner. Turner cites introductory mathematics, science, and writing courses as ones where students benefit from “learning by doing,” she says. 

2. Where might AI get in the way of students learning? 

Students, says Turner, need to build their own confidence and intuition so that they can use AI effectively. 

“Many instructors think it is crucial for students to be able to do some parts of the process before they use AI,” she says. “For example, students may need to do reading on their own rather than turning to AI summaries in introductory classes, because that’s where they learn the basics of reading for academic work.” 

It may even be the case, she adds, that AI use does not support an instructor’s learning goals. “Faculty want students to see the value of their own ideas and thoughts rather than have students rely on AI tools,” Turner says. 

3. Where does AI use really help students learn? 

For Turner, AI is most effective when it supports students’ ability to ask questions and reflect on their understanding. She points to the value of chatbots in empowering students to work through questions on their own—like, for example, if they’re studying at 3 a.m. and can’t expect a quick response from an instructor or teaching assistant. These bots—answer bots that are trained on course information, or tutor bots that don’t give answers but instead prompt questions that deepen their thinking—are designed to be supportive, not act as shortcuts. 

“Tutor bots in particular can help students by scaffolding the learning so the student can come to the answer,” Turner says. 

In some cases, she adds, these AI-powered supports can help students meet more rigorous goals than they could otherwise. One component of planning AI use in assignments and activities, Turner explains, is to think about how they promote reflection. She cites the activity example of Elizabeth Emery, an associate professor of nutrition science who asked students to use a large-language model (LLM) to generate a meal plan. She then asked students to evaluate where the LLM made mistakes or could have been better. 

“Then students can start to figure out, ‘If I do a first draft with AI, what are the next steps I have to take to make sure it’s an effective plan?’” says Turner. “It allows them to do the work much more quickly but also allows them to reflect on what questions are that they should always be asking themselves.” 

Reflection, says Turner, is the key ingredient in learning, especially when working with AI. 

CETLI’s AI Month panel 

Panelists for the upcoming AI Month event include:

• Faizan Alawi, a professor of pathology at Penn Dental Medicine
• Elizabeth Emery, a practice associate professor at Penn Nursing
• Bhuvnesh Jain, the Walter H. and Lenore C. Annenberg Professor in the Natural Sciences, at SAS 
• James Petersson, a professor of chemistry at SAS · Sarah Pierce, the Denise A. Rotko Associate Dean for Legal Practice skills at Penn Carey Law
• Nat Trask, an associate professor of mechanical engineering and applied Mechanics at Penn Engineering 
• Sarah Pierce, Denise A. Rotko Associate Dean for Legal Practice at Penn Carey Law

Turner says panelists were chosen as a representation of different kinds of AI activities. Some, like Emery’s, demonstrate the value of reflection, while others, like Petersson’s, showcase how to think about AI as a tool to answer questions. Following the panel discussion, attendees can participate in an informal reception to talk more with the panelists. 

“I’m excited to get people who are interested in AI in conversation with faculty members who are already using it,” says Turner. “I think having people from all ends of the spectrum together is going to be a useful way for people to think about what AI can do for their classes, and where to avoid AI.”