Artificial Intelligence: A Hands-On Lesson in Responsible Integration in Schools

During the Tang Institute’s second Lunch and Discussion event of the year, Nicholas Zufelt, instructor in math and computer science at Phillips Academy, guided a conversation about the nuanced relationship between students, teachers, and generative AI in the classroom. As a teacher and a Tang Fellow actively engaged in the ethi{CS} project, Zufelt is deeply involved in the subject on a daily basis. His expertise in the field made the session a worthy one.

Zufelt rooted his talk in a recent research paper by Dr. Ethan Mollick of The Wharton School and Dr. Lilach Mollick of Wharton Interactive. The paper explored innovative ways students could interact with AI beyond the pitfalls of a passive ​“write this paper for me” approach — an approach that rightfully scares many classroom teachers who are concerned about AI and academic integrity. Using the paper as a springboard, Zufelt offered a roadmap for a more interactive and collaborative relationship with this evolving technology. Teachers could, he suggested, have the AI present and play the roles in a simulation for students to work through or have it be a teammate in group work.

To demonstrate to students how teachers might feel if the students have AI do their work for them, Zufelt turned the tables. He proposed that teachers could use AI to write assessments of students. He then read a prompt he had written for a [fictional] end-of-term instructor comment, and AI generated a seemingly impressive response. However, when he asked the students how they would feel about teachers using AI for this purpose, a sea of thumbs down conveyed a sense of betrayal. They would feel cheated.

Zufelt then asked, ​“Why be a student?”

The resounding answer?

“To learn.”

This exchange led Zufelt to emphasize that AI should not replace the learning process but, rather, enhance it. The key, he pointed out, lies in understanding the goal — learning — and how AI can be a partner, not a replacement, in the journey of acquiring knowledge. As education consultant Eric Hudson shared a few weeks ago during his Lunch & Discussion, AI should augment human capabilities rather than diminish them.

One of the most important pieces of Zufelt’s talk were the eight essential rules for responsible AI use in education, starting with the crucial step of always getting permission from teachers before using AI and culminating with a powerful reminder: ​“Always work hard — it should never replace your thinking!”

1. Get permission from your teacher
2. Be playful; experiment
3. Give it examples to emulate; ask it for quantity
4. Don’t stress about the prompt, instead offer feedback (as you would a person)
5. Be resilient; AI messes it up sometimes
6. Be skeptical; AI ​“hallucinates”
7. Maintain privacy
8. Always work hard
   

Finally, Zufelt had attendees partner up and tackle a few scenarios he’d prepared using the Mollick paper. For twenty minutes, attendees explored AI’s potential applications as a student, a simulator, and a tutor. It was a powerful exercise for the students, teachers, and staff in the room, and for many, a first-time experience.

Before breaking, Zufelt reiterated that responsible integration of AI in education will require thoughtful consideration, open communication, and a commitment to preserving the essence of the learning experience. He encouraged students to be both skeptical and curious, urging them not to accept everything AI presents but also to approach the technology with a sense of exploration and fun.

On Monday, December 4, Zufelt will be presenting on similar work at Loomis Chaffee’s AI Symposium ​“Navigating Uncertainty, Embracing Possibility: AI in Independent Schools.”

_____
Nicholas Zufelt is an instructor in mathematics, statistics, and computer science at Phillips Academy. He is also a Tang Fellow, working on the ethi{CS} project.